text
string | predicted_class
string | confidence
float16 |
|---|---|---|
However, the biological insights (the dynamic ranges of metabolite abundance in a biological system, conservation relationships in networks of interrelated compounds, etc.) and the currently inherent analytical limitations (suboptimal quality of extraction, detection and ionization capabilities, etc.) , call for caution in handling noise elimination and redundant signals in metabolomic datasets. The effort to produce a noise-free and non-redundant data matrix (post-processing) could result in a loss of information, some of which could be informative to comprehensively assess metabolic pathways for a better understanding and description of the regulatory mechanisms underlying the global biological responses . Furthermore, as recently demonstrated, some of the ion peaks that could be regarded as a source of redundancy (e.g., adduct formation) might be very crucial and actually needed in metabolite annotation and differentiation .
|
study
| 99.94 |
The created matrices (Table 1) were then imported into SIMCA (Soft Independent Modeling of Class Analogy) version 14 software (Umetrics, Umea, Sweden) for statistical modeling: principal component analysis (PCA) and orthogonal projection to latent structures discriminant analysis (OPLS-DA) (generally used approaches in metabolomics data analysis for data overview/descriptive exploration, and explicative/predictive analysis, respectively). Total variation in metabolomic data is multifactorial, comprising of the sum of biological variation (induced and non-induced) and technical variation . Therefore such data, with its inherent properties and structures (including large number of variables, nonlinearity, heteroscedasticity, missing values), imperatively requires special attention during statistical handling to avoid a risk of model overfitting , manipulation, and confusion of statistical findings and distortion of the results, which can lead to incorrect data interpretation and false discovery .
|
study
| 99.94 |
Before performing PCA and OPLS-DA, the data was mean-centered (to put all variables on equal footing) and Pareto-scaled (to adjust for measurement errors as to have homoscedasticity in the data). There are different methods for dealing with missing values, and each method/approach impacts on downstream statistical analyses . In the present study, the SIMCA software uses an adjusted nonlinear iterative partial least squares (NIPALS) algorithm (with a correction factor of 3.0) in handling the missing values. The threshold of missing values is, by default, 50%, and in the four matrices (from the same raw data, Table 1) no observations or variables had missing values exceeding the permitted tolerance. Thus, to assess the effect of processing parameters (mass tolerance and intensity threshold, Table 1) on the statistical output, the quality and characteristics of computed PCA and OPLS-DA models were comparatively evaluated. The statistically-extracted discriminating variables from the four scenarios (Table 1) were also compared.
|
study
| 100.0 |
For principal component (PC) analyses, the results showed that varying the processing parameters (mass tolerance and intensity threshold) affected the maximum (suitable) number of computed PCs, optimized using seven-fold cross-validation, to explain the variation in data X: changing from a five-PCs to a six-PCs models (Table 2). Notably, only the “R1” significant components, i.e., those producing an increase in Q2, were retained. Although, visually, the sample clustering in the PCA scores space (constructed from the first two PCs) show no significant difference across the four datasets (Figure 1A,B and Figure S1A,B), the model quality was clearly affected. This can be assessed by inspecting the PCA parameters and diagnostic tools, which are computed and displayed graphically or numerically. In computing a PC model, strong and moderate outliers (observations that are extreme or do not fit the model) are often formed. Strong outliers have a high leverage on the model, shifting it significantly and reducing the predictability, whereas the moderate outliers correspond to the temporary perturbations (in the process/study), indicating a shift in the process/study behavior .
|
study
| 100.0 |
Strong outliers are identified from scores and Hotelling’s T2 range plots. The latter is a multivariate generalization of Student’s t-test, providing a check for observation adhering to multivariate normality . When used in conjunction with a scores plot, the Hotelling’s T2 defines the normality area corresponding to 95% confidence in this study. Inspecting the scores and Hotelling’s T2 range plots for the calculated four PC models (Figure 1A,B and Figure S2), no strong outliers were observed. The moderate outliers, on the other hand, are identified by inspecting the model residuals (X-variation that was not captured by the PC model). The detection tool for the moderate outliers is the distance to the model in X-space (DModX), with a maximum tolerable distance (Dcrit) .
|
study
| 100.0 |
In this study, for all four datasets (Table 1), the DModX was normalized in units of standard deviation, with the significance level of 0.05. Inspecting the DModX plots (Figure 1C,D and Figure S1C,D) showed the existence of some moderate outliers. What is important to notice is that these moderate outliers were different in the four PC models (Figure 1C,D and Figure S1C,D), suggesting that varying of the two processing parameters (mass tolerance and intensity threshold) clearly altered the structure in the X-space (particularly in higher-order components), and this impacts the statistical description thereafter. The moderate outliers were further investigated by computing the contribution plots (Figure S3) and no sample/observation had variable(s) with critical deviation from the rest of the dataset.
|
study
| 100.0 |
Furthermore, the model fit (R2X) and predictive power (Q2) diagnostic parameters were evaluated for the computed four PC models. The model fit informs how well the data of the training set can be mathematically reproduced indicating, quantitatively, the goodness of fit for the computed model. The R2X, thus, quantitatively describes the explained variation in the modeled X-space . The predictive ability of the model, on the other hand, was estimated using cross-validation, providing a quantitative measure of the predicted variation in X-space. A change in data processing parameters (mass tolerance and intensity threshold) clearly affected PCA, altering the model quality. The positive change in both mass tolerance and intensity threshold parameters resulted in an increase in R2X and Q2, with a substantial difference observed in the predicted variation, Q2 (Table 2). These results demonstrate that the upstream metabolomic data processing and treatment affect the outcome of the statistical analyses, which then would impact, both quantitatively and qualitatively, the mining of “what the data says” .
|
study
| 100.0 |
For supervised multivariate analyses (OPLS-DA in this case), the calculated models were validated and assessed. Firstly, to note that for all the OPLS-DA models of the four datasets, there was clear discrimination between the sample groups in the scores space (Figure 2A and Figure S4). The analysis of variance testing of cross-validated predictive residuals (CV-ANOVA) was used to assess the reliability of the obtained models . The computed OPLS-DA models for the four datasets, to separate multivariate relationships into predictive and orthogonal variation, were statistically good models with p-values significantly lower than 0.05 (Table 2). Furthermore, the response permutation test (with n = 50) was used to validate the predictive capability of the computed OPLS-DA models. In this statistical test the R2 and Q2 values of the true model are compared with that of the permutated model. The test is carried out by randomly assigning to the two different groups, after which the OPLS-DA models are fitted to each permutated class variable. The R2 and Q2 values are then computed for the permutated models and compared to the values of the true models .
|
study
| 100.0 |
The results indicate that the calculated models have much higher R2 and Q2 values (Figure 2B and Table 2) and, thus, the computed true OPLS-DA models are statistically far better than the 50 permutated models for each dataset. Assessing the total variation in X-space (predictive and orthogonal) explained by the models, the results show that the R2X values were different: a change in mass tolerance and intensity threshold affect the amount of variation explained by the computed models (Table 2). For variable selection, the OPLS-DA loading S-plots were evaluated (Figure 2C). This loading plot has an S-shape provided the data are centered/Pareto-scaled, and aids in identifying variables which differ between groups (discriminating variables), i.e., variables situated at the upper right or lower left sections in the S-plot. The p1-axis describes the influence of each X-variable on the group separation (modeled covariation), and the p(corr)1-axis represents the reliability of each X-variable for accomplishing the group separation (modeled correlation). Variables that combine high model influence (high covariation/magnitude) with high reliability (i.e., smaller risk for spurious correlation) are statistically relevant as possible discriminating variables : |p| ≥ 0.05 and |p(corr)| ≥ 0.5 in this study.
|
study
| 100.0 |
Furthermore, to avoid variable selection bias , the significance of the variables from the loading S-plot was assessed using, firstly, the variable importance in projection (VIP) plot (SIMCA 14 software). The latter summarizes the importance of the variables both to explain X and to correlate to Y. The higher the VIP value (exceeding 1.0) the more significant is the variable in the complex analysis in comparing the difference between two or more groups . Each selected variable from the S-plot (with high model influence and reliability, and VIP score >> 1.0) was further evaluated using a dot plot (Figure 2D). The latter is similar to a histogram and kernel density estimation (but algorithmically different), computing each observation as a unit: the observations are sorted into “bins” representing variable sub-ranges . A very strong discriminating variable has no overlap between groups (Figure 2D). The mathematical description of the mentioned algorithms and methods (e.g., VIP, dot plot, etc.) is beyond the scope of this paper; the reader is, thus, referred to the cited literature.
|
study
| 99.9 |
The statistically-validated discriminating variables from each model (representing each dataset generated from varying the two mentioned processing parameters) were then compared. The results demonstrate that the change in pre-processing parameters (mass tolerance and intensity threshold, in this case) affected the downstream statistical analyses, particularly the statistically-selected variables: comparing these variables showed some overlap, but also each method had unique variables (Figure 3). This observation compliments and corroborates the above PCA results that data processing and treatment (prior to statistical analyses) alter not only the infographics, but also the extracted information, which might impact the interpretation thereafter.
|
study
| 100.0 |
To also evaluate the effect of data (pre)-treatment algorithms on downstream chemometric models, different scaling and transformation methods were applied on the data matrix created using Method 1 (Section 3.1, Table 1). The scaling methods explored are center (Ctr), autoscaling (also termed unit variance (UV)), and Pareto, and the transformation methods used were logarithmic and power transformation (the formulae are described in the experimental section). To avoid confusion, the definitions for scaling and transformation methods used are those in the SIMCA manual (User’s Guide to SIMCA 13, 2012), which are also related to descriptions found in the cited literature .
|
study
| 100.0 |
Following scaling and transformation, PCA and OPLS-DA models were constructed/fitted and evaluated to assess the influence of these data pre-treatment methods on the models’ quality, classification accuracy, feature selection/extraction, and the subsequent interpretability of the data. The data pre-treatment is an essential step in the metabolomic data analysis pipeline as it enables the preparation of the data for downstream analyses, minimizing variable redundancy and making all variables more comparable in size .
|
study
| 99.94 |
For PC analyses, the results showed that scaling and transformation methods affected the dimension of the PC-space optimized using seven-fold cross-validation, the sample clustering in the PCA scores space (e.g., constructed from the first two PCs), and the moderate outliers detected in the DModX plots (Figure 4 and Figure S5). Furthermore, the metric used to assess the model fit (or explained variation) and predictive ability of the computed PCA models were R2 and Q2 . The inspection of these diagnostic metrics shows that scaling and/or transformation remarkably affected the amount of explained variation (the goodness of fit) by the model and its predictive ability (Table 3).
|
study
| 100.0 |
As in Section 2.1, for supervised multivariate analyses (OPLS-DA in this case), the calculated models were validated, and the robustness and reliability of the models assessed. In addition to the R2 and Q2 metrics, the CV-ANOVA was used to assess the reliability of the obtained models and the response permutation test (with n = 50) was used to validate the predictive capability of the computed OPLS-DA models . Furthermore, in both Section 2.1 and Section 2.2, predictive testing was also employed to assess the best pre-processing and pre-treatment workflow (Figures S6 and S7). The results tabulated in Table 3 demonstrate that the scaling and transformation methods affected significantly not only the explained variation R2 (both predictive and orthogonal) but also the classification accuracy, reliability, predictive capability of the model and, subsequently, extracted variables (Figure 5). The supervised learning models computed following for instance UV-scaling and/or log-transformation (particularly in this case), would not be chemometrically/statistically trusted as the classification of these models could be by chance, as indicated by the permutation validation tests (lower R2 values compared to the permutated models, Table 3).
|
study
| 100.0 |
These results (Table 3) demonstrate that the choice of pre-treatment method(s) is crucial and may depend not only on the biological information to be acquired but also on the data analysis method to follow. For instance, a chemometric approach/method that focuses on (dis)similarities would differ from the PCA attempting to explain as much variation as possible in as few components as possible. Thus, varying data properties using data pre-treatment methods may, for instance, enhance the results of a clustering method, while obscuring the results of a PCA model .
|
study
| 99.94 |
These results (varying pre-processing parameters and pre-treatment methods) clearly demonstrate that handling the multivariate data from untargeted metabolomic analyses is indeed challenging. Both Figure 3 and Figure 5 depict that the data analysis outcome of an untargeted metabolomic data is remarkably influenced by the upstream data handling, such as pre-processing and pre-treatment methods, and the algorithms applied. The symbiosis of chemometrics and metabolomics is illustrated here, with a clear demonstration that an understanding of data structures and data analysis methodologies is mandatory for stepping forward from data to information as comprehensively as possible.
|
study
| 96.4 |
Furthermore, Figure 3 and Figure 5 then raise questions with regard to what could be the best methodological approach for data pre-processing and pre-treatment, given an untargeted metabolomics dataset (as it is in this case) that exhaustively explores the data. Could these results (including the tabulated models’ description in Table 2 and Table 3) actually be pointing to the problem of “multiplicity of good models”/“multitude of descriptions” (Rashomon Effect) ? Or is it, indeed, an indication that, currently, no single chemometric method can actually extract all of the information from an untargeted metabolomic dataset. Different chemometric methodologies and algorithms are constantly being developed to cope with systems biology-perspective demands , but maybe the chemometric principle of “largest variance is most important” might not hold true in all cases, as the total variation in the metabolomic data is multifactorial. On the other hand, to extract relevant information may require searching the “needle in the haystack” methods .
|
review
| 63.78 |
There is certainly a “symphony” of data analysis strategies and approaches throughout the pipeline, from post-acquisition steps to the variable selection. Algorithms or methods used at each step of the pipeline affect the downstream analyses and outcome . In this study, using the same LC-MS raw data, but changing the pre-processing parameters (Section 3.1) and data pre-treatment methods (Section 3.2) affected the downstream statistical outcome, thus illustrating (to a certain extent) not only that the metabolomic data are indeed information-rich, but also the limitation in existing chemometric methods and the need of uttermost care in data handling. Furthermore, the results demonstrate that, currently, the possible “Archimedean” methodical point for an optimal extraction of information from untargeted metabolomic datasets could be the exploration of existing chemometric and statistical methods at each step of data analysis pipeline.
|
study
| 99.94 |
However, these observations and generalizations are built on an assumption that the study design and data quality assurance are always correctly executed and considered , and the underlying philosophy of chemometrics is efficiently applied throughout the metabolomics study from the start to biological interpretation . To maximize the value of metabolomic data, exploration of different algorithms and methods might be the best trade-off currently. As this study demonstrates, different steps involved in data mining are interdependent (Figure 6), and methods employed in each step would influence the downstream steps. However, although an exploration of different algorithms and methods is encouraged, this should be guided and discerned based on a thorough manual examination of the raw data and a strong analytical and chemometric knowledge.
|
study
| 99.94 |
Raw data from an untargeted ultra-high performance liquid chromatography-mass spectrometry (UHPLC-MS) metabolomics study, in this case of sorghum plants responding dynamically to infection by a fungal pathogen, Colletotrichum sublineolum, was employed. Briefly, two groups of samples were used in this study: fungal-infected (treated) and control (non-infected) samples, labelled T and C, respectively. The control group consisted of nine samples, whereas the treated group consisted of 15 samples. Each sample was a methanol extract from 10 plants. Analytical data of methanol-based plant extracts were acquired in both positive and negative centroid ion mode; but for this paper, only the positive data were further processed. The m/z range was 100–1000 Da and the data were acquired by applying a Waters Acquity UHPLC system coupled in tandem to a Waters photodiode array detector and an electrospray SYNAPT G1 HDMS Q-TOF mass spectrometer (Waters, Milford, MA, USA), applying a method as previously described . A lock spray source was used allowing online mass correction to obtain high mass accuracy of analytes. Leucine encephalin, [M + H]+ = 556.2766 and [M − H]− = 554.2615, was used as the lock mass, being continuously sampled every 15 s, thus producing an average intensity of 350 counts scan−1 in centroid mode. By using a lock mass spray as a reference and continuously switching between sample and reference, the MassLynxTM software can automatically correct the centroid mass values in the sample for small deviations from the exact mass measurement.
|
study
| 100.0 |
Quality control (QC) pooled samples were used to condition the LC-MS analytical system so as to assess the reliability and reproducibility of the analysis, and for non-linear signal correction . Sample acquisition was randomized and the QC sample (six injections) was analyzed every 10 injections to monitor and correct changes in the instrument response, with each sample being injected three times. Furthermore, six QC runs were performed at the beginning and end of the batch to ensure system equilibration. Such sample randomization provides stochastic stratification in sample acquisition so as to minimize measurement bias. In the PCA space, the QC samples were clustered closely to each other (results not shown, as it is not the focus of this study), thus confirming the stability of the LC-MS system used, the reliability, and reproducibility of the analysis.
|
study
| 100.0 |
Visualization and data processing were performed using MassLynx XSTM 4.1 software (Waters Corporation, Manchester, UK). Only the centroid electrospray ionization (ESI) positive raw data were used in this study. The MarkerLynxTM application manager of the MassLynx software was used for data pre-processing (matrix creation). Four dataset matrices (hereafter referred to as Methods) were created by changing mass tolerance and intensity threshold settings: Method 1 (mass tolerance of 0.005 Da and intensity threshold of 10 counts), Method 2 (mass tolerance of 0.005 Da and intensity threshold of 100 counts), Method 3 (mass tolerance of 0.01 Da and intensity threshold of 10 counts), and Method 4 (mass tolerance of 0.01 Da and intensity threshold of 100 counts). For all of the Methods, the parameters of the MarkerLynxTM application were set to analyze the 1–15 min retention time (Rt) range of the mass chromatogram, mass range 100–1000 Da, and alignment of peaks across samples within the range of ±0.05 Da and ±0.20 min mass and Rt windows, respectively.
|
study
| 100.0 |
The MarkerLynxTM application uses the patented ApexTrack (termed also ApexPeakTrack) algorithm to perform accurate peak detection and alignment. MarkerLynxTM initially determines the regions of interest in the m/z domain based on mass accuracy (mass tolerance). The ApexTrack algorithm controls peak detection by peak width (peak width at 5% height) and baseline threshold (peak-to-peak baseline ratio) parameters. In this study, these parameters were calculated automatically by MarkerLynxTM. The ApexTrack also calculates the baseline noise level using the slope of inflection points. Thus, for peak detection, the ApexTrack algorithm consists of taking the second derivative of a chromatogram and locates the inflection points, the local minima, and peak apex for each peak, to decide the peak area and height. A “corrected” Rt is then assigned and the data are correctly aligned, with the alignment of peaks across samples within the range of user-defined mass and Rt windows. Following the peak detection, the associated ions are analyzed (the maximum intensity, its Rt and exact m/z mass) and captured for all samples.
|
study
| 100.0 |
An additional data cleaning step, a peak removal step denoted by user-defined peak intensity threshold (and noise elimination level) parameter, is conjugated to the alignment algorithm: briefly, if a peak is above threshold in one sample and if it is lower than the threshold in another sample it lowers the threshold for that sample until it reaches the noise elimination level. The noise is understood as residual peaks in the background (from electronics, nebulizer gas, solvents, cleanliness of source, column, etc.) and/or below the noise elimination threshold. MarkerLynxTM also performs data normalization. In this study normalization was done by using total ion intensities of each defined peak. Prior to calculating intensities, the software performs a patented modified Savitzky-Golay smoothing and integration.
|
study
| 99.94 |
Although parameters, such as mass tolerance and intensity threshold (which define the real peak versus noise), can be regarded as relatively instrument-dependent (or a property of acquired data), they can be changed (within certain limits): mass tolerance can be set to the mass accuracy of the acquired data (which was 4.9 mDa in this study) and twice this value; hence, in this study mass tolerance was varied within these limits (0.005 and 0.01 Da). Considering the complexity and high-dimensionality of the samples (particularly in plant metabolomics), and considering the mathematical limitations of chemometric algorithms, it is essential to explore the processing methods (combinations of sets of parameters: selected as objectively and optimally as possible) so as to maximize the mining of the raw data.
|
study
| 100.0 |
The MarkerLynxTM-generated data matrices were exported into SIMCA software, version 14 (Umetrics, Umea, Sweden) for statistical analyses. An unsupervised method, principal component analysis (PCA), and a supervised modeling, orthogonal projection to latent structures-discriminant analysis (OPLS-DA), were employed. The data pre-treatment methods used included scaling and transformation. These two types of data pre-treatment were explored as described in Section 2.2. The scaling methods looked at were center (Ctr), autoscaling, (also known as unit variance, UV) and Pareto, and the transformation methods used were logarithmic and power transformation. The formulae (or mathematical description of these methods) can be found in the cited literature and in the SIMCA version 13 manual (User’s Guide to SIMCA 13, 2012). In this study, the logarithmic transformation was 10Log (C1 × X + C2) where C1 = 1 and C2 = 0; and the power transformation was (C1 × X + C2) C3 where C1 = 1, C2 = 0, and C3 = 2. As described in the results, the computed models were validated.
|
study
| 100.0 |
Using the same raw dataset, and exploring and applying different methods and algorithms in handling the data, the study clearly demonstrates how crucial the data pre-processing and pre-treatment steps are in a metabolomics data analysis pipeline. These steps significantly affect the chemometric models computed downstream, including the variation explained by the models, the classification accuracy and the quality of the models. However, the inferred observations from this study are limited, as being drawn from a single dataset (from plant samples). Applying the same exercise to different datasets from other sample types (e.g., cell extracts, biofluids, etc.) might provide more insights and subsequently a formulation of generalized guidelines. Thus, it suffices here to point out that an understanding of the data structures and the approach adopted for handling a specific untargeted metabolomic dataset will definitely influence the data analysis outcome, as demonstrated in this study. Stepping forward from untargeted metabolomic data, to information, and finally knowledge, is not a trivial endeavor or a simple “syllogistic” approach.
|
study
| 99.94 |
As encouraged by the Metabolomics Society , a proper and detailed reporting of the data analysis methodology used in a metabolomic study is essential and “ethically” obligatory, for clarity of the etiology of the inferences of the study and ascertaining the reproducibility of the latter. Furthermore, the growing call for submission of metabolomic studies (and raw data) into the repositories, such as MetaboLights , is to be encouraged, as further mining of datasets (with different data analysis scopes) can yield more information and more biological insights . Untargeted metabolomic studies, in general, generate large amounts of data that are exceedingly rich in information and, consequently, realistically challenging to mine, interpret, and pursue mechanistically in a comprehensive biological context. Furthermore, even though the current study used only one method for peak detection (and varying its parameters), it should be noted that there are a wide variety of workflows (vendor-specific, commercial software, and freeware) available for peak picking/detection. Since various algorithms are used by these different peak detection methods, the choice thereof can have a significant influence on the processed end results of a study . Hence, careful and thoughtful use of pre-processing and processing tools is mandatory to be able to make a critical judgment on the outcomes following those two essential steps in the metabolomics workflows. Furthermore, exploration of data analysis methods (as demonstrated in this study) and data sharing (via data repositories) are encouraged so as to maximize the value of these metabolomic datasets.
|
review
| 99.7 |
Global ecosystems are currently exposed to unprecedented environmental change. Although global change is often associated with range contractions or even extinctions (Thomas et al. 2004), some species emerge as profiteers and expand their range into newly emerging habitat (Le Roux and McGeoch 2008). Such expansions are usually interpreted as responses to changing environmental conditions at the range edge (Hickling et al. 2006). Moreover, they could simply be the result of phenotypic plasticity allowing species to exploit changing environmental conditions (Gienapp et al. 2008; Merilä and Hendry 2014) or could even be affected by bacterial infections (Goodacre et al. 2009). However, the potential for a successful expansion is not necessary inherent to a species, but could also be acquired by evolutionary adaptation (Parmesan 2006; Hill et al. 2011). While environmental change could initially trigger range expansions, only the interplay with adaptation to, for example, novel climatic conditions at the range edge may lead to massive expansions as currently observed for many species (Clements and Ditommaso 2011; Franks and Hoffmann 2012). An expanding population might initially be released from selection pressure by competitors, parasites, and predators. This release has been shown in many invasive species and might speed up the population's regrowth and facilitate an adaptation to novel environmental conditions in a new habitat (Keane and Crawley 2002; Menéndez et al. 2008; Phillips et al. 2010). After the evolution of novel phenotypes, for example, increased physiological tolerance, a population can further expand its range, much farther than by environmental change alone (Clements and Ditommaso 2011; Franks and Hoffmann 2012). Indeed, several examples of rapid adaptive divergence during contemporary range expansions and biological invasions have been observed (Huey et al. 2000; Phillips et al. 2006; Colautti and Barrett 2013; Krehenwinkel et al. 2015). Even in the face of considerable ongoing gene flow between expanding and native populations, strong selection can mediate adaptations (Saint‐Laurent et al. 2003). Gene flow and secondary contact might even provide expanding populations with novel variation and facilitate adaptive differentiation (Rius and Darling 2014). In fact, genetic signatures of isolation of populations by environmental conditions appear to be quite frequent and often override simple isolation by distance (Sexton et al. 2014). This suggests an important role of ecological divergence during contemporary range expansions. However, the interplay of different environmental factors, phenotypic plasticity, gene flow, selection, and adaptation remains poorly understood, especially in the initial stages of contemporary range expansions.
|
review
| 99.56 |
Spiders provide some ideal model species to study the contribution of these factors to successful range expansions. Many spiders are efficient long‐distance dispersers. By passive wind‐mediated dispersal on silk threads, so called ballooning, they are capable of moving large distances in a single generation (Foelix 2010). They can thus quickly follow changing environmental conditions as evidenced by several ongoing poleward range expansions (Hänggi and Bolzern 2006; Muster et al. 2008; Kumschick et al. 2011). Recent work also shows that such range expansion can be accompanied and facilitated by rapid adaptive divergence of expanding lineages (Krehenwinkel and Tautz 2013). Spiders are also subject to intense parasitism by hymenopterans and dipterans. By causing more than 50% mortality in some populations, these parasites may impose strong selection (Finch 2005). A reduction of parasite load in novel environments could then result in rapid population growth.
|
study
| 99.94 |
The yellow sac spider, Cheiracanthium punctorium (Villers, 1789), offers an interesting example of an expanding spider species. It is a comparably large, free ranging, and nocturnal hunter, living on fallows and meadows. Their historical European distribution was largely limited to warm climate regions in southern Europe with the southern Upper Rhine Plain marking the northern limit of its distributional range (Bellmann 2006). Beginning in the mid‐20th century, isolated occurrences have been reported from more northern latitudes in the eastern German plains and as far north as Sweden (Muster et al. 2008). These novel occurrences were geographically restricted to small, isolated sites. However, over the last decade, a marked expansion has been noted from these focal areas and the species is now found in very high abundances in most of northeastern Germany, southern Sweden, and the Baltic States. At the same time, large gaps are present between these sites, in which the spider is currently not found, for example, Poland, northwestern Germany, and Denmark (Muster et al. 2008; Jonsson 2005; Krehenwinkel, pers. observation, Talvi and Rutkowski pers. communication). The spread of C. punctorium is of special relevance, as it is one of very few Central European spider species whose venom can cause noteworthy effects on humans and may require medical treatment (Muster et al. 2008).
|
study
| 74.25 |
Here, we aim to identify factors associated with the expansion success of the species in Europe. In particular, we hypothesize that genetic structure might reflect a pattern of environmental isolation between native and expanding populations, possibly as a result of rapid adaptation. Moreover, we hypothesize that a release from brood parasitism might be associated with range expansion success. Finally, we predict the species' potential distribution using species distribution modeling, highlighting its future expansion in Europe.
|
study
| 99.94 |
To test these hypotheses, we analyzed 400 C. punctorium specimens from several native and two invaded locations throughout the species' European range. We used a set of microsatellite loci, nuclear 28S rDNA, and mitochondrial COI sequences, to infer genetic structure and signatures of isolation. We also employed morphological measurements to show phenotypic differentiation of lineages. In addition, we estimated the frequency of eggsac parasitism in selected native and historical expanding populations to gather insights in the role of parasite release during the range expansion. Finally, we use multivariate comparisons between the realized niches of native and expanding populations to predict the species' historical and current potential distribution and to identify climatic factors associated with genetic differentiation of lineages and expansion success.
|
study
| 100.0 |
We include 399 samples spanning most of the European range of C. punctorium. Our sampling of the native range covers southwestern Russia, eastern Ukraine, Slovenia, southwestern Germany, the Iberian Peninsula, and Italy. These sites comprise the historical native range of C. punctorium. In addition, we analyze specimens from two expanding foci: one in the Baltic States Latvia and Estonia and the other in the northeastern German state Brandenburg. Expanding populations were identified based on recent reports of novel occurrences or massive expansion from previously known, isolated occurrences (Muster et al. 2008; Jonsson 2005; Krehenwinkel pers. obs.; Tallvi, pers. comm.). The Baltic populations have only been recently identified and have probably been established less than a decade ago. Brandenburg constitutes an older population, which was first recorded in the mid‐20th century (Muster et al. 2008; Jonsson 2005; Krehenwinkel pers. obs.). Specimens were sampled by hand or sweep net. Additional samples were acquired from the Senckenberg Museum of Natural History. All mature female specimens were examined under a Leica dissecting microscope, and two morphological measurements were taken using a Leica measuring eyepiece (Leica, Wetzlar, Germany): prosoma width (at its widest part) as a measure of body size, and the length of the first leg's longest segment, the femur. As leg length relates to body size, we used relative leg length as ratio of femur length and prosoma width. We took the morphological measurements for 381 specimens from the Iberian Peninsula, Italy, Slovenia, Russia, northeastern Germany, and the Baltic states. While it is easy to quantify, body size is a trait of major ecological importance (Gardner et al. 2011). Moreover, body size has been shown to rapidly evolve in range expanding arthropods (Huey et al. 2000; Krehenwinkel and Tautz 2013).
|
study
| 99.94 |
We compiled all available occurrence data for the species in Europe from the GBIF database (Dooley 2002), classical and online publications, a survey among European arachnologists, and by own field studies. In total, our distribution dataset comprises 228 records. Due to its notoriety as a venomous species, the occurrence and expansion of C. punctorium is quite well documented and new occurrences are usually accompanied by media coverage (Muster et al. 2008). Our approach thus probably allowed a reasonably accurate approximation of the species' contemporary range. To identify ecological differences between old and recently established populations, we distinguished native and expanding populations. Seventy‐three occurrences could be identified as native, 136 as expanding, while 19 occurrences had uncertain status. To distinguish native and expanding populations, we performed two separate analyses. A first analysis was performed with a reduced dataset of populations, for which we could acquire genotyping results (see Molecular analysis). Population subdivision was based on genetic differentiation. Based on these results, we performed a second analysis, assigning populations to the native and expanding groups based on proximity to genotyped populations and considering the species' documented expansion. Expanding sites were assigned to those, where a recent expansion of the species has been reported, or where it was only recently discovered. The results of both analyses were compared, to ensure a proper population assignment.
|
study
| 99.94 |
Based on GPS coordinates, we estimated the potential distribution of the species in Europe using the hypervolume package in R (R Development Core Team 2008; Blonder et al. 2014; Blonder 2015). We used 19 gridded bioclimatic variables with a spatial resolution of 2.5 arc min (approximately 4 km within Europe) representing average climate conditions between 1950 and 2000 from the WorldClim database (Hijmans et al. 2005). These 19 variables were transformed into four principal components with eigenvalues >1, which were used as orthogonal niche axes to compute the hypervolumes (for factor loadings, see Fig. S2).
|
study
| 100.0 |
The hypervolume analysis estimates the realized niche space of a species based on a random sample derived from multivariate kernel density estimates of the species' occurrence records in (Blonder et al. 2014). Two different approaches for the delimitation of the respective hypervolume are available: the first approach uses a bandwidth (herein = 1) enclosing the species records in environmental space (termed BWD hereafter), whereas the second approach delimits the realized niche space based on a multivariate minimum convex polygon (termed MMCP hereafter, see Supplementary Material S4 for a visualization of both approaches). Based on the hypervolumes of two taxa (herein native and expanding populations, as well as a set of records which could not be assigned to either of them), it is possible to compute various statistics including the volume of the realized niches per taxon, the intersection of two taxa as well as the unique volumes, the centroid distance between two taxa, and the Sorensen index [i.e., for hypervolumes A and B: S = 2*|A int B| / (|A| + |B|), Blonder 2015]. Furthermore, an inclusion test allows a projection of the hypervolumes into geographic space, representing the potential distribution of the species based on the estimates of the realized niches.
|
study
| 99.94 |
One leg of each specimen was removed with sterile forceps, and genomic DNA extracted using the Qiagen DNeasy kit, according to the manufacturer's protocol (Qiagen, Hilden, Germany). A total of nine Iberian and nine Baltic DNA samples were chosen each from three populations, and their concentration measured on a NanoDrop Fluorospectrometer (NanoDrop, Wilmington, DE, USA). Equal DNA amounts were pooled for the two geographic regions, and each pool sequenced in 300 bp paired reads on one flow cell of an Illumina MiSeq. Library preparation and sequencing were conducted according to the manufacturer's protocols (Illumina, San Diego, CA, USA). The resulting DNA reads were quality trimmed with a minimum quality of 20 using Popoolation (Kofler et al. 2011). A de novo assembly was then generated including both libraries and using CLC genomic workbench with a word size of 45, a bubble size of 98, a minimum contig length of 1000 and including a scaffolding step (CLC Bio, Boston, MA, USA). We identified microsatellite markers in the genome assembly using Tandem Repeats Finder (Benson 1999). We then established a set of 14 polymorphic markers, which were genotyped for 260 specimens from all studied populations. Marker isolation, primer design, PCRs and genotyping were performed as described in Krehenwinkel and Tautz (2013). Microsatellite alleles were called using GeneMapper (Applied Biosystems, Foster City, CA, USA). Microsatellite analyzer (MSA) was used to generate population distance matrices (Nei's genetic distance) with 500 bootstrap replicates for the microsatellite data and to estimate genetic diversity (Dieringer and Schlötterer 2003). A neighbor‐joining tree was calculated based on the distance data with Phylip (Felsenstein 2002) and edited in MEGA 6. Pairwise F ST values between populations were calculated in Genepop on the web (Raymond and Rousset 2002). In addition to the distance based analysis, we performed genetic clustering analysis using STRUCTURE (Pritchard et al. 2000; Falush et al. 2003). STRUCTURE was run with an admixture model, k‐values from 1–10, 10 replicates per k‐value, a burnin period length of 100 000 and 100 000 MCMC replications after burnin. The number of populations was estimated with STRUCTURE HARVESTER (Earl and von Holdt 2011) using the approach described in Evanno et al. (2005). We performed two independent STRUCTURE analyses: one with all sampled populations and one with only native European populations, to exclude potential masking effects by highly divergent expanding populations.
|
study
| 100.0 |
Approximately 900 bp fragments of the 28S rDNA (for 105 specimens) and the mitochondrial COI gene (for 399 specimens) were PCR amplified using the Qiagen Multiplex PCR kit according to the manufacturer's protocol and using the primers 28Srd4.8a and 28Srd7b1 (Schwendinger and Giribet 2005) and C1‐J‐1718‐spider and C1‐N‐2776‐spider (Vink et al. 2005). While COI is well known as a useful phylogeographic and taxonomic marker in spiders (Krehenwinkel and Tautz 2013), the more slowly evolving 28S might also be suited to uncover interspecific divergence in spiders (Jäger et al. 2015). We thus use it as a backup for the COI data here. Sequence traces were edited using Codon Code Aligner (Codon Code Corporation, Centerville, MA, USA) and aligned in MEGA 6 (Tamura et al. 2013) under default alignment parameters. The mitochondrial sequences were translated to amino acids to detect potential presence of pseudogenes.
|
study
| 100.0 |
Next, we aimed to identify those climatic factors, which directly relate to genetic divergence and thus possibly contribute to divergence or cohesion of lineages across Europe. The association of the 19 bioclimatic variables to genetic differentiation was evaluated using GESTE (Foll and Gaggiotti 2006). In addition to the bioclimatic variables, we included longitude and latitude for all populations in the analysis, to test for isolation by distance. GESTE was run with 10 pilot runs and a sample size of 10 000, with a thinning interval of 20 and a burnin of 50 000.
|
study
| 100.0 |
After mating, C. punctorium females build silken retreats on top of grass stalks, in which they construct a single eggsac and guard the hatching offspring till they die in late autumn. The guarding female will attack intruders ferociously (Bellmann 2006). Nevertheless, we found numerous eggsacs to be infected by larvae of a parasitic dipteran that consume all spider eggs before hatching. Infection with these parasites thus leads to a complete reproductive failure of the female. We screened nine native populations in Italy, and ten expanding populations in the Baltic States and northeastern Germany for the presence and abundance of this parasite. Moreover, we gathered information from arachnologists working in the respective areas. To screen for parasites, 10–20 female retreats per site were opened and eggsacs inspected for parasitic larvae. Some larvae were taken to the laboratory where COI barcodes were generated for identification of the species using standard arthropod barcoding primers (Folmer et al. 1994). PCR and sequencing were conducted as described above. The resulting sequences were blasted against the GenBank database, to identify related sequences.
|
study
| 100.0 |
The MiSeq run yielded ~14*106 sequences for the Iberian DNA pool and ~20*106 for the Baltic one. The genome assembly amounted to 148 601 contigs in ~ 236*106 bases, with an N50 of 1425 of which we could isolate primers for 4453 microsatellite loci. We genotyped populations across the European range of the yellow sac spider (Fig. 1A) for 14 of these loci. This revealed significant genetic structure. Our STRUCTURE analysis suggested k = 3 populations. We found a homogeneous native European population, including samples from the Iberian Peninsula, Italy, Slovenia, southwestern Germany and Russia (Fig. 1B). In contrast to the homogeneity in the native range, we identified two genetically separated expanding populations: one in the Baltic States and another in the northeastern German state of Brandenburg. The northeastern German populations were less divergent in comparison with the native cluster than the Baltic ones (allele frequency difference Baltic versus. Native = 0.045, allele frequency difference NE‐Germany versus Native = 0.022). Moreover, the two expanding clusters were clearly divergent (allele frequency difference Baltic versus NE‐Germany = 0.034). A second STRUCTURE analysis excluding the two expanding populations suggests additional and finer‐scaled genetic structure in the native European cluster with k = 2. Eastern and western European populations from Russia and the Iberian Peninsula each form separate clusters. The eastern cluster extends to the Balkans and southwestern German, while populations from Italy appear admixed between these two clusters (Figure S2). These results were generally supported by an allele sharing phylogeny of the data with the expanding population from northeastern Germany and the Baltic States forming well‐separated clades. The phylogeny also suggested a similar differentiation of Russian and Iberian populations within the native range, with other native populations grouping between these two (Figure S3). An analysis of F ST differentiation between populations additionally confirmed our previous analyses. The two expanding populations were divergent from each other and all native populations, while the differentiation between native populations was considerably lower (Table 1).
|
study
| 100.0 |
(A) Sampling map for the analyzed yellow sac spider populations, showing native populations in green, expanding northeastern German ones in blue and expanding Baltic ones in red. B. Result of a STRUCTURE analysis assuming K = 3 and based on 14 microsatellite loci for the according populations. Colors correspond to those in Fig. 1A.
|
other
| 98.1 |
We did not find a pronounced difference of genetic diversity between native and expanding spider populations. This held true for allelic richness (AR) and expected heterozygosity (EH). The highest diversity was found in Russia (AR = 3.69, EH = 0.63), followed by Iberia (AR = 3.46, EH = 0.60) and Italy, then northeastern Germany (AR = 3.31 3.62, EH = 0.58), and finally, the lowest in the Baltic States (AR = 2.85, EH = 0.51). The only significant difference in diversity was found for the Baltic populations, which showed a reduction of allelic richness and expected heterozygosity compared to all other populations (anova, Games–Howell post hoc test, P < 0.01). Russian, Iberian, and populations from northeastern Germany did not show significant differences in diversity.
|
study
| 100.0 |
To summarize, our genetic analyses identified a fairly homogenous native European population from Russia over Slovenia and Italy to southern Germany and the Iberian Peninsula. Only slight differentiation between the eastern and western native range is evident. This was contrasted by two divergent expanding populations in northeastern Germany and the Baltic states. We did not find a consistent signature of reduced genetic variation in recently established populations compared to native ones.
|
study
| 99.94 |
The mitochondrial haplotype network (Fig. 2) showed a dumbbell pattern of two divergent haplogroups, separated by 18 substitutions (groups A and B). Haplogroup A showed several closely related starburst patterns of haplotypic radiations, with most haplotypes divergent by 1–3 substitutions. In contrast, haplogroup B was quite homogeneous and consisted of only four haplotypes. Haplogroup A was found throughout the spider's range. Native European populations from the Iberian Peninsula, Italy, Slovenia and southwestern Germany almost all exclusively carried haplotypes of group A. The distribution of haplotypes within these native populations did not show pronounced geographic structure. A single Iberian specimen carried a haplotype of group B. This same haplotype was found at a much higher frequency in Russian populations (about 30%). Considering the high divergence between haplogroups A and B, Russian populations showed a higher nucleotide diversity than other native ones (π Russia = 0.0122, π Iberia = 0.0036, π Italy = 0.0038), while haplotype diversity was more comparable (HDRussia = 0.841, HDIberia = 0.679, HDItaly = 0.818). Similar to the Russian populations, about 30% of the specimens from the Baltic States also carried haplogroup B. Their haplotype was distinct from the Russian one by a single substitution. Their nucleotide diversity was high, while their haplotype diversity was reduced compared to native populations (π Baltic = 0.0115, HDBaltic= 0.617). An even more pronounced reduction of diversity was found for expanding populations in northeastern Germany, which carried only two slightly divergent haplotypes from group A (π NE‐Germany = 0.0002, HDNE‐Germany = 0.189). In contrast to the mitochondrial data, the 28S rDNA was completely monomorphic across all of Europe, with all 105 sequenced specimens sharing the same 28S alleles.
|
study
| 100.0 |
Median‐joining mitochondrial haplotype network for ~850 bp of the COI gene in 399 specimens from native and expanding European yellow sac spider populations. Unless indicated by numbers, branches between haplotypes imply single substitutions. White circles represent hypothetical, unobserved interconnections between haplotypes.
|
study
| 55.62 |
Our body size measurements indicated a homogenous native population from Russia (average body size ± standard deviation = 4.72 ± 0.59 mm), Italy (4.82 ± 0.51 mm) and the Iberian Peninsula (4.62 ± 0.41 mm), as well as two isolated expanding ones in the Baltic states (4.06 ± 0.47 mm) and northeastern Germany (4.42 ± 0.50 mm) (Fig. 3A). Body size, as measured by the prosoma width, of native populations was not significantly different, while the expanding populations were each distinguished by a significantly smaller body size from the native populations (anova, Games–Howell post hoc test, P < 0.05). In addition, we found a significant size difference between the two expanding populations, with specimens from northeastern Germany being larger than those from the Baltic. Overall, the body size data were highly congruent with our microsatellite analysis.
|
study
| 100.0 |
(A) Average body size (prosoma width) of mature females in native and expanding European yellow sac spider populations. (B) Relative leg length (femur I length / prosoma width) of mature females in native and expanding European yellow sac spider populations. Bars represent the 95% confidence interval of the mean; colors correspond to those in the sampling map in Fig. 1A.
|
study
| 99.8 |
The relative leg length showed slightly different results (Fig. 3B). Here, we could distinguish a group comprising western specimens from Iberia (average relative leg length ± standard deviation = 1.35 ± 0.06) and Italy (1.37 ± 0.07), which were distinct from other populations by significantly shorter legs (anova, Games–Howell post hoc test, P < 0.05). Russian spiders and those from the two recently colonized locations shared a very similar relative leg length (average relative leg length Russia = 1.45 ± 0.06, Baltic = 1.45 ± 0.08, northeastern Germany = 1.45 ± 0.09). This analysis thus corroborated a closer relationship of Russian and expanding populations.
|
study
| 100.0 |
DNA barcoding identified the eggsac parasite as Sarcophaga sexpunctata (Fabricius, 1805), a widely distributed flesh fly known to parasitize eggsacs of different spider species. We screened 10 expanding and nine native populations for S. sexpunctata infections. Of nine screened native Italian populations, all showed infections. We estimated the abundance of S. sexpunctata infections in four of these populations. Between 30% and 90% of all eggsacs were parasitized, on average 50%. At the same time, none of the 10 analyzed expanding populations from the Baltic States and northeastern Germany were affected by brood parasitism. Inquiries with fellow arachnologists did not reveal any incidence of parasitism in additional expanding populations in Sweden and northeastern Germany (Jonson and Friman pers. comm.).
|
study
| 100.0 |
Principal component analysis revealed four PCs with eigenvalues >1. PC1 was predominantly composed of temperature and precipitation‐related variables (Bio1: Annual Mean Temperature, Bio6: Minimum Temperature of the Coldest Month, Bio9: Mean Temperature of the Driest Quarter, Bio11: Mean Temperature of the Coldest Quarter, Bio13: Precipitation of Wettest Month, and Bio16: Precipitation of Wettest Quarter). PC2 was most strongly correlated with Bio14: Precipitation of the Driest Month and Bio18: Precipitation of the Warmest Quarter, while PC3 was driven by Bio2: Mean Diurnal Temperature Range, Bio4: Temperature Seasonality, and Bio7: Temperature Annual Range. PC4 represented mainly ‘Precipitation Seasonality’ (Bio15). For all factor loadings, see Supplementary Material S4.
|
study
| 100.0 |
Based on a comprehensive set of species records, the realized climatic niche space of the native populations was much larger than the niche space occupied by the expanding populations and those whose status could not be determined (Volume BWD Native = 3172.6; Volume BWD Expanding = 19.7; Volume BWD Unknown = 5.3; Volume MMCP Native = 722.3; Volume MMCP Expanding = 236.2; Volume MMCP Unknown = 108.2). For visualization, see Fig. 4 and Supplementary Material S4. The shared volumes between native and expanding populations were comparatively small (Intersection BWD Native / Expanding = 57.0; Intersection MMCP Native / Expanding = 6.4) leading to low–very low Soerensen indices (S BWD Native / Expanding = 0.12; S MMCP Native / Expanding = 0.004). This indicates a strong niche shift between native and expanding populations, wherein the combination of PC1 and PC3 had the highest contribution to the differentiation (Fig. 4).
|
study
| 100.0 |
(A) Potential current distribution of yellow sac spiders in Europe, based on only native European populations as well as populations with unknown status. (B). Potential current distribution of yellow sac spiders in Europe, based on all currently known occurrences of expanding populations, as well as populations with unknown status. (C, D) Estimated hypervolumes based on a bandwidth approach (BWD) and multivariate minimum convex polygons (MMCP). For high‐resolution versions of the biplots, see Supplementary Material S4.
|
study
| 85.94 |
The potential distributions derived from the hypervolumes indicated a much wider distribution for the species than currently known. This holds especially true for northern European expanding populations. While the species is currently only known from isolated foci, our model indicates a potential distribution over Poland, northern Germany, and the countries around the southern Baltic Sea (Fig. 4).
|
other
| 96.4 |
An association of genetic structure with bioclimatic variables and longitude and latitude was tested using GESTE. The highest probability model of our GESTE analysis suggests an association of only Bio10 (Mean Temperature of Warmest Quarter) with genetic structure over Europe (posterior probability = 0.228, marginal probability = 0.568, α = −0.63). For the analyzed populations, the genetic differentiation decreases with increasing summer temperatures (Fig. S1). In contrast, neither geographic distance nor winter cold explain the observed genetic differentiation.
|
study
| 100.0 |
Contemporary range expansions are often simply attributed to global change (Hickling et al. 2006), but instead might be caused by a complex interplay of factors. A combination of global change, ecological processes, phenotypic plasticity, and finally adaptation might pave the way for the currently observed massive expansions of many species (Gienapp et al. 2008; Franks and Hoffmann 2012). This might also hold true for the yellow sac spider. Warming climate and the increasing availability of fallow land could have initially enabled an expansion into novel habitats. The expansion did not occur on a broad front. Instead, several suitable patches of habitat were colonized outside of the native range. First isolated occurrences of the yellow sac spider were in fact reported from northeastern Germany in the mid‐20th century (Muster et al. 2008). Like many other spiders, C. punctorium probably possesses high dispersal ability, due to passive, wind‐mediated ballooning of young spiderlings (Bell et al. 2005; Krehenwinkel et al. 2016). The range wide homogeneity of 28S rDNA and overall low microsatellite differentiation within the native range also suggests connectivity of populations by gene flow. Suitable emerging habitat and changing temperatures can probably be quickly reached by the species.
|
study
| 99.94 |
Our genetic analyses suggest two independent colonization events: one into northeastern Germany and one into the Baltic States. Our genetic and morphological analyses suggest similarity between Russian and both expanding populations. An important source of expanding populations might be located in eastern European steppe. The climate in this region is distinguished by cold winters and thus more comparable to that in the newly invaded C. punctorium habitats. The invading populations could therefore be preadapted to the novel conditions, which might facilitate their initial establishment.
|
study
| 99.9 |
The observation of two divergent mitochondrial haplotypes in the Russian and Baltic expanding populations resemble recent findings in Argiope bruennichi, another expanding spider species (Krehenwinkel and Tautz 2013; Krehenwinkel et al. 2015). Here, an admixture of two formerly separated lineages relates with rapid divergence of expanding and native populations. This finding is in line with many recent studies that associate expansion success and genetic admixture (Kolbe et al. 2004; Nolte and Tautz 2010; Rius and Darling 2014). Interestingly, our microsatellite analysis suggests the contact of a western and eastern European native genetic cluster in Central Europe. Expanding populations might thus have been founded from admixed populations. However, a deeper geographic sampling and more genetic data will be necessary to quantify an association of admixture and expansion.
|
study
| 100.0 |
While the source of expanding populations might be preadapted to cold winters, expanding spiders also have to cope with cooler and more humid summers in northern Europe. The environmental isolation of expanding populations could be explained by divergent selection between the newly colonized and native environments. The time lag from the initial establishment in the mid‐20th century to the current expansion might be explained by the necessary waiting time for adaptations to emerge (Clements and Ditommaso 2011). Such lag phases are commonly observed in biological invasions and range expansions (Sexton et al. 2009). An adaptation to novel climatic regimes has been shown to evolve quickly in spiders, even in the face of high gene flow (Tanaka 1996; Krehenwinkel and Tautz 2013). Apart from direct climate tolerance, a particular phenotype in northern European populations could be found in their reduced body size. Yellow sac spiders are an annual species. They grow from spring till summer and then reproduce. A reduced body size might thus be an adaptation to a shorter growing season in cooler climates and a necessity for earlier reproduction in more northern latitudes. Latitudinal size gradients have been described for many arthropod species and can evolve quickly (Huey et al. 2000; Krehenwinkel and Tautz 2013). However, more experimental data, for example, from reciprocal transplant experiments, common garden studies and whole genome sequencing, will be necessary to support the assumption of an adaptive divergence of the studied populations. Based on our data, we currently cannot exclude an involvement of phenotypic plasticity in the observed differences. Most phenotypes are known to be affected by environmental context, and plasticity plays an important role in species responses to contemporary climate change (Gienapp et al. 2008; Merilä and Hendry 2014).
|
study
| 99.94 |
The species' range expansion might be facilitated by a release from brood parasitism in the recently colonized range (Menéndez et al. 2008; Phillips et al. 2010). By effectively removing 50% of each generation in native populations, Sarcophaga sexpunctata imposes a very strong mortality factor. Interestingly, S. sexpunctata is widely distributed in Europe and feeds on eggs of different spider species (Povolný 2000), yet, according to our results, does not parasitize on the expanding populations of C. punctorium. Northern European S. sexpunctata populations might require more time to adopt C. punctorium as a novel food source. Moreover, S. sexpunctata has been considered thermophilic and might show a decreasing population density with increasing latitude (Povolný et al. 2003).
|
study
| 100.0 |
Yellow sac spiders are currently massively spreading over northern Europe. Our species distribution model suggests a much wider potential distribution of C. punctorium in northern Europe than currently observed. We predict a colonization of most of Germany, the Baltic States, Poland, Denmark, and southern Sweden. The yellow sac spider is the only venomous Central European spider whose bites frequently necessitate medical treatment. A yellow sac spider bite can be very painful and might even cause side effects such as nausea, vertigo, fever, and shivers (Sacher 1990; Weimann et al. 2011; Papini 2012; Nentwig et al. 2013). Although the C. punctorium venom is not life threatening and its bites are rare even in densely colonized areas (Sacher 1990), our predictions are nevertheless important, in that they highlight one of the consequences of global change. In addition to increasing reports of disease vectors expanding their ranges (Khasnis and Nettleman 2005), another problematic outcome of global change is the spread of venomous species into new territories.
|
study
| 99.94 |
Our study provides evidence for rapid genetic and ecological differentiation of expanding populations during a contemporary range expansion. This differentiation might be caused by adaptive divergence, which could additionally fuel the success of the expansion. However, more experimental data (e.g., from reciprocal transplants or genome sequencing) will be required to rule out phenotypic plasticity as the driving force of the observed expansion. In the coming decades, an unprecedented expansion of the venomous yellow sac spider into northern Europe is expected, highlighting one of the health‐related perils of the ongoing climatic changes.
|
study
| 99.9 |
The following data are available online from: GenBank: COI sequences and 28S rDNA sequences Dryad (http://dx.doi.org/10.5061/dryad.2q8p6): Georeferenced sampling sites and records used for the distribution modelCOI and 28S rDNA sequence alignmentsMicrosatellite data and Microsatellite primersMorphological measurements for all specimensGenome assembly.
|
other
| 99.9 |
Figure S4. A. Potential current distribution of yellow sac spiders in Europe, based on only genotyped native and expansive European populations. B. Estimated hypervolumes based on a bandwidth approach (BWD) computed only for genotyped native and expansive European populations. C. Estimated hypervolumes based on multivariate minimum convex polygons (MMCP) computed only for genotyped native and expansive European population.
|
other
| 99.7 |
The ocular surface is an integrated unit comprising corneal and conjunctival epithelia, meibomian glands (MGs), main and accessory lachrymal glands, and trigeminal neurons; their dysfunction results in a scarce or unstable tear film that causes dry eye, with a higher incidence among postmenopausal women .
|
other
| 99.9 |
Dry eye disease (DED), as a multifactorial disease, presents a complex aetiology and pathophysiology . It is widely known that systemic and metabolic dysfunctions, such as lower intake of omega-3 and omega-6 fatty acids, menopause, acne and ovarian dysfunction, diabetes mellitus, and use of systemic medications (antihistamines, b-blockers, antidiuretics, and antidepressants) could favour the development of DED and increase the severity of the disease . Sex hormone imbalance plays a crucial role in the pathophysiology of different ocular surface diseases including dry eye, with a different impact of oestrogens and steroids. In detail, their imbalance may significantly increase the risk and modify the course of DED, since serum oestrogen levels are strongly associated with the development and progression of dry eye . This is also supported by the fact that women are more likely to experience DED during periods of substantial hormonal alteration, such as pregnancy, lactation, oral contraceptive use and after the menopause .
|
review
| 99.9 |
Experimental and human studies have demonstrated that androgen levels are essential for the normal lacrimal gland function and for structural organisation, and that prolactin and oestrogens play important roles in providing an adequate hormonal environment for optimal tear production . In fact, it was demonstrated that systemic replacement treatment with combined esterified oestrogen and methyltestosterone may be efficacious in treating dry eye syndrome of various aetiologies . Moreover, receptors for androgens, oestrogens, progesterone and prolactin have been identified in several ocular tissues, including the main lacrimal gland and MGs , leading to the hypothesis that tear steroids may have a key role in the physiology of these glands.
|
review
| 99.56 |
Since tears have already revealed important insights for studying eye disorders , and considering our experience in the development of liquid chromatography-tandem mass spectrometry (LC-MS/MS) methods for the determination of metabolites related to pathologies , we aimed to measure steroid levels directly on tear samples, for the first time. Thus, a robust, specific, and selective method was developedfor the simultaneous quantification of the following steroids: cortisol (CORT); corticosterone (CCONE); 11-deoxycortisol (11-DECOL); 4-androstene-3,17-dione (ADIONE); testosterone (TESTO); 17α-hydroxyprogesterone (17-OHP); and progesterone (PROG). The proposal methodology was applied for the analysis of a small casuistry of tear samples from 14 female dry eye patients and 13 female controls.
|
study
| 100.0 |
Following the criteria described above limit of detection (LOD) and lower limit of quantification (LLOQ) were established to be: 0.30 and 0.50 ng/mL for CORT; 0.1 and 0.29 ng/mL for CCONE; 0.05 and 0.10 ng/mL for 11-DECOL; 0.05 and 0.08 ng/mL for ADIONE; 0.02 and 0.08 ng/mL for TESTO; 0.05 and 0.08 ng/mL for 17-OHP. The only exception was PROG, for which LOD and LLOQ coincided with the lowest concentration Schirmer strip calibrator (0.05 ng/mL). The correlation between steroid concentration and signal intensities, including two blank Schirmer strip calibrators, was linear: from 0.50 to 20.80 ng/mL for CORT; from 0.10 to 1.00 ng/mL for 11-DECOL; from 0.08 to 2.91 ng/mL for ADIONE; from 0.08 to 1.00 ng/mL for TESTO; from 0.08 to 4.20 ng/mL for 17-OHP; and from 0.05 to 4.20 ng/mL for PROG. For CCONE the linearity was observed from 0.29 to 4.03 ng/mL, but the method resulted to be not linear for this steroid. CCONE has the same nominal molecular mass as 11-DECOL, and even if the chromathographic conditions were able to chromatographically resolve CCONE and 11-DECOL (Figure 1), the signal for CCONE after LC-MS/MS analysis of Schirmer strip samples was found to be low, reflecting the typical low endogenous concentration of this steroid. Table 1 shows the investigated linear ranges, the calculated calibration functions, and the corresponding correlation coefficients (r2). The r2 for all analytes was >0.993 over their concentration ranges, and all calculated concentrations for the calibrators were within ±10% of the assigned values. The intra- and inter-assay coefficient of variation (CV) and deviation from assigned values was ≤15.2% and ≤13.9% for all analytes, respectively (Table 2) . The mean analytical recovery from the extraction at each Quality Control (QC) level for all steroids, without internal standard correction, was also summarized in Table 2. Recovery using analyte peak areas exclusive of internal standard normalization gives an indication of the efficiency of analyte extraction resulting from the chosen method of sample preparation. According to the principle of isotope dilution internal standardization, the addition of isotopically labelled internal standards for each analyte prior to the extraction should compensate for any variation in ionisation efficiency due to residual matrix effects and for losses in recovery of the extraction process .
|
study
| 100.0 |
The LC-MS/MS method we developed was employed for the analysis of tear samples collected on Schirmer paper strips in a total of 27 people divided in 14 DED female patients and 13 healthy control females. Since about the 25% of tears samples had steroid concentration below our defined LLOQ, with the exception for CORT and PROG, and considering the limited number of patients for this first application, we decided to express our data in term of response (analyte peak area/internal standard peak area) for all the seven steroids.
|
study
| 100.0 |
In order to define data distribution in each group, data matrix was statistically processed performing D’Agostino and Pearson omnibus normality test. Once normality was accepted, Student’s t-test was performed, otherwise the Mann Whitney test was carried out to assess the significantly expressed variables between the groups. Response levels were measured for all the seven steroids. CORT, ADIONE, and 17-OHP response levels resulted significantly decreased (p-value < 0.05) in dry eye patient samples compared to control samples. CCONE response level also resulted decreased in this comparison, although for this steroid we did not achieve the optimal conditions after the validation assay procedure. Figure 2 shows tear steroid profiling of dry eye and in healthy control patients (panel A). Data are mean values and bars represent the corresponding standard deviations (SD): ** indicates p < 0.01, and *** indicates p < 0.001. In Figure 2, panel B–D represent tear sample distribution for CORT, ADIONE and 17-OHP response level, respectively. In order to evaluate the diagnostic predictive power, we performed receiver operating characteristic (ROC) curve based on the three significantly different steroids obtained. After this investigation, the area under the curve (AUC) was 0.964, highlighting good sensitivity and specificity in discriminating the two study groups (Figure 3, panel A). We also performed the 100 cross-validation showing the predicted class probabilities of each sample, underlying good predictivity of the proposed model (p < 0.01) in discriminating dry eye patients (class 1) from healthy people (class 0) (Figure 3, panel B,C) .
|
study
| 100.0 |
DED describes a group of tear film disorders that cause irritation and ocular surface damage. The most common subtypes of DED include aqueous and lipid deficiency, although most patients with DED have abnormalities in both tear components . Considering dry eye prevalence and cost, known risk, and underlying pathogenesis of DED seems to be determinant to achieve effective prevention and treatments. Age (menopause) and female sex are considered as valid risk factors for DED development as stated in the literature . In recent years, LC-MS/MS has had a pivotal role in routine clinical chemistry thanks to a wide range of applicability, easy sample preparation, and high analytical specificity, which allows for the identification, characterization and quantification of chemical compounds as target analytes based on their respective molecular masses and fragmentation patterns . This technique allows the simultaneous analysis of a large number of metabolites from many different biological matrices in order to investigate multifactorial disease, as in the case of DED.
|
review
| 99.9 |
Even if different LC-MS/MS methods for the determination of a number of steroids from serum, plasma, dry-blood spot and urine samples has been already described, in this work for the first time we present a robust and sensitive LC-MS/MS method that allows us to analyse seven tear steroids by performing only one extraction procedure from tears collected on Schirmer strip . We have demonstrated good sensitivity since we are able to detect up to 0.3 ng/mL for CORT, 0.05 ng/mL for 11-DECOL, ADIONE, 17-OHP and PROG, and 0.02 ng/mL for TESTO in tear samples. It can be an advantage since to date, the clinical used immunoassay for serum sex hormones have not been able to reliably measure the low concentrations, especially regarding TESTO in females and children . It is also to consider that usually in clinical practice, it is needed one assay for each steroid of interest, while LC-MS/MS methods give the opportunity to monitor multiple analytes in a single analysis. Therefore, to test the seven steroids in tears by using the conventional immunoassay methods it would be necessary to perform seven different sample collection. The methodology described involves a simple extraction of target steroids (CORT, CCONE, 11-DECOL, ADIONE, TESTO, 17-OHP and PROG) from tears collected on Schirmer strip to obtain a specific profile through the multiple reaction monitoring (MRM) acquisition mode for each analyte using two transition (quantifier and qualifier ions) to ensure lack of interferences. Thus, our LC-MS/MS method proved to have good specificity, sensitivity, and also good linearity and intra- and inter-assay imprecision. The use of LC-MS/MS allowed us to obtain chromatographic resolution of the analytes of interest, even for CCONE and 11-DECOL, which have the same nominal molecular mass. Unfortunately linearity for CCONE was not achieved.
|
study
| 99.94 |
Many association of DED metabolic traits have been already reported, in particular it has been highlighted a relationship between DED and serum steroids. Vehof et al. found a strong significant association between DED and decreased levels of some serum androgens, suggesting the use of androgens as a potential pathway for the treatment of DED . In fact, Bizzarro et al. proved that patients with DED and Sjögren syndrome showed a significant increase of tear volume in the Schirmer test and a decrease in the Bijsterveld score after oral testosterone undecanoate treatment with respect to placebo patients . A correlation between DED and serum steroids has been observed, but it is unknown if it is possible to confirm the same relationship with the tear steroids. This feature is exactly the starting point of our pilot study.
|
study
| 99.94 |
Following a careful development and validation of an LC-MS/MS method for the measure of seven steroids in tears collected on paper strips, a small cohort of DED patients and controls was recruited and analysed for the determination of a tear steroid profiling potentially related to DED.
|
study
| 100.0 |
Statistical investigations of our data showed a significant reduction in tear response levels for CORT, ADIONE and 17-OHP in DED patients when compared to healthy controls. As also reported by Vehof et al. , a significant association was found for DED and decreased levels of some androgens. It is worth to mention that androgen levels should influence the structure and function of the lacrimal and Meibomian gland, with decreased androgen levels leading to lower tear volume, reduced tear film stability through decreased quality and quantity of meibomian gland lipids, decreased tear turnover rate, and hyperosmolarity . ADIONE and 17-OHP are both secreted into the blood circulation by the adrenal glands . Since ADIONE and 17-OHP tear levels resulted to be decreased in the case of DED, it is interesting to note that a positive correlation was demonstrated between these two steroids by Rudnicka et al. . Moreover, CORT, a glucocorticoid, is produced by adrenal glands, in particular in the zona fasciculata of the adrenal cortex. Since CORT is one of the end product of steroidogenesis and derived by 17-OHP, we may speculate about a possible link between the decreased 17-OHP tear levels and the lower tear levels observed for CORT in DED patients. Interestingly, sex hormone PROG showed an opposite trend, even if not significant, increasing in dry eyes patients compared to controls. Higher levels of PROG is consistent with literature data that suggests a greater DED incidence during pregnancy, lactation, and oral contraceptive use .
|
study
| 99.94 |
In summary, this work revealed for the first time that tears represent a precious source of information for the study of the pathophysiologic state of the eye. In particular, in this pilot study steroid tear levels were measured and correlated to DED. Even if only a small cohort of tear samples were analysed, our innovative LC-MS/MS application added evidence of steroid level alteration in the tears of DED patients. Since the exact relationship between serum and tear steroid levels remains unknown, it could be interesting to study these two biological fluids from a large cohort of DED patients.
|
study
| 100.0 |
Dry eye patients and controls had to show a best correct visual acuity of 8/10, a mean intra-ocular pressure lower than 18 mm Hg, a central cornea thickness ranging from 530 to 570 μm, normal dilated funduscopy. At the moment of enrollment, patients with DED were not on therapy and did not receive topical steroids during the last 2 months. Exclusion criteria were diabetes mellitus, ocular lymphoma, sarcoidosis, autoimmune deficiency syndrome, corneal dystrophy, and non-DED-related ocular surface inflammatory diseases, systemic or topical therapy potential affecting the corneal status, glaucoma, topical therapy with steroids or nonsteroidal anti-inflammatory drugs, use of contact lenses, and previous ocular surgery. Exclusion criteria for normal controls were history of systemic or topical therapy, ocular or systemic diseases in the previous 12 months, pregnancy, and contact lens wear. Both of the eyes were evaluated, but one eye per subject was randomly chosen (using a computer-generated random number list) for statistical analysis.
|
study
| 99.94 |
All tear samples were collected at Opthalmic Clinic of University “G d’Annunzio” of Chieti-Pescara between September 2016 and February 2017. Tear samples from healthy and dry eye patients were collected on graduated strip for dry eye testing (Schirmer test I). Dry eye was diagnosed according to the International Dry Eye Workshop criteria . Schirmer test I results were expressed as the length of the strip that was wet after 5 min. The Schirmer strips were purchased from EasyOpht (Busto Arsizio, VA, Italy). Once folded, the Schirmer paper strip at the mark, tears were collected asking the patients to look up and pulling the lower lid gently downward. After 5 min, the strip was removed from the eye and the length of the moistened area was measured using the millimeter scale on the strip. Then the filter paper was placed in a 2.0 mL Eppendorf tube, left dry at room temperature and stored at −80 °C. The Schirmer tests were considered as positive for dry eye diagnosis when strips were wet below 10 millimeters in five minutes. In case of the strip becomes completely wet before the indicated time, it may be removed in advance. This procedure was performed contemporarily for right and left eye for each patient. To preserve anonymity, full drug and medical histories were not available other than dry eye diagnosis and treatment. For this pilot study 13 healthy female and 14 dry eye female patients were analysed (from 25 to 73 years old), with three samples collected for each individual. In particular 10 patients presented an untreated dry eye without a diagnosis of autoimmune diseases, 2 patients were affected by dry eye thyroidism related and 1 patient affected by Sjogren disease (new diagnosis). More than 80% of patients with DED presented with a severity level 2, according to criteria previously reported . All subjects gave their informed consent for inclusion before they participated in the study. The study was conducted in accordance with the Declaration of Helsinki.
|
study
| 99.94 |
CORT, CCONE, 11-DECOL, ADIONE, TESTO, 17-OHP and PROG were purchased from Sigma-Aldrich® (St. Louis, MO, USA). 2H3-CORT, 2H8-CCONE, 2H5-11-DECOL, 2H5-ADIONE, 2H5-TESTO, 2H8-17-OHP, 2H9-PROG were from CHSTM MSMS Steroids Kit, PerkinElmer® (Turku, Finland). Water (H2O), methanol (MeOH), acetonitrile (ACN) LC-MS grade were from Romil-Pure chemistry® (Cambridge, UK). Formic acid and ethanol LC-MS grade were from Sigma-Aldrich® (St. Louis, MO, USA). HPLC solvent additive was from CHS MSMS Steroids Tool Box, PerkinElmer® (Turku, Finland).
|
other
| 99.9 |
Each endogenous steroid (1 mg) was dissolved in ethanol (stock solution) and stored at −20 °C. Stock solutions of each compound were diluted in methanol/water 50:50 with a final concentration of 10 µg/mL (tuning solution) and stored at −20 °C. The internal standard mixture (IS mix) from PerkinElmer® kit was reconstituted with 1.25 mL of ACN and stored at −20 °C. Before the extraction procedure, the Daily Precipitation Solution (DPS) containing Internal Standards was prepared by diluting 1:1000 the IS mix in two consecutive steps with ACN with 0.1% formic acid. After diluting the tuning solution, we obtained the working solutions for each steroid of interest as follow: CORT 50 ng/mL, CCONE, TESTO, 17-OHP and PROG 5 ng/mL, 11-DECOL and ADIONE 10 ng/mL. Calibrators were prepared from each working solution to achieve final concentrations as reported in Table 3. The same procedure was used to obtain QC samples that, considering the small concentration range of interest, resulted to coincide with L2, L4 and L6 calibrator levels (Table 3).
|
study
| 100.0 |
Calibrators and QCs in Schirmer paper strips were prepared wetting each paper strip with 90 µL of calibrator and QC solutions for each levels of calibration. Calibrators, QCs, tear samples collected on Schirmer strip were cut into 2–3 mm paper pieces and transferred into 2.0 mL microcentrifuge tube (Eppendorf®, Hamburg, Germany), paying attention to wash the required equipment with MeOH before each sample preparation. After adding 200 μL of DPS containing IS, each sample was gently mixed (20 °C, 15 min) in a Thermomixer (Eppendorf®) and then centrifuged (4210 rcf, 20 °C, 30 min). The organic layer (135 µL) was transferred into a new 1.5 mL tube and dried in a SpeedVac for 30 min. The residue was then reconstituted with 90 μL of H2O/MeOH 60:40, gently mixed in a Thermomixer (20 °C, 15 min), briefly centrifuged and finally transferred into polypropylene vial (Waters Corporation, Milford, MA, USA). The vials were capped, gently mixed, and placed in the system autosampler for analysis.
|
study
| 99.94 |
The LC-MS/MS system was a HPLC Alliance HT 2795 Separations Module coupled to Quattro Ultima Pt ESI tandem quadrupole mass spectrometer (Waters Corporation, Milford, MA, USA). The system operated in positive electrospray ionization mode using MassLynx v4.1 software (Waters). For HPLC analysis, the Luna® 3 μm C8, 100 Å, 50 × 3 mm column with Security Guard Cartridges, C8, 4.0 × 2.0 mm as guard column (Phenomenex®, USA) was used. Autosampler injections of 50 μL were made into sample loop in full mode. The mobile phase comprised a binary solvent system: 98% H2O and 2% MeOH (Solvent A) and 100% MeOH (Solvent B), both containing 0.025% of solvent additive. The initial solvent composition was 65% A and 35% B. The mobile phase gradient profile involved three steps: increasing from the initial conditions to 45% B within 4.0 min and then to 65% B within 9.0 min holding for 0.5 min before reaching 100% B until the end of the analysis. The total run time was 18.00 min, injection to injection. The flow rate was 0.35 mL/min and the column was maintained at 45 °C.
|
study
| 99.9 |
The mass spectrometer ionization source settings were optimized for maximum precursor ion yields for each steroid. This was achieved by injecting the tuning solution for each individual compound, all the parameters are summarized in Table 4. Two mass transitions were optimized for each analyte, with a single transition being used to monitor the corresponding deuterated internal standards. The first transition was used to quantify the target analyte and the second to qualify the identity of the target compound using a confirmatory ion ratio. The capillary voltage was 3.5 kV, source temperature was 120 °C, desolvation temperature was 400 °C, and the collision cell gas pressure was 3.5 × 10−3 mbar argon. The inter-channel and inter-scan delay times were 0.02 and 0.1 s, respectively. The dwell time was 0.45 s for CORT, 0.30 for CCONE and ADIONE, 0.25 for 11-DECOL, TESTO and 17-OHP, 0.35 for PROG. The same dwell time parameters as the ones of the endogenous compounds were used for the respective internal standards. Functions 1–7 (Table 4) refer to the MRM experiments created for each analyte.
|
study
| 100.0 |
Data processing and quantification were performed using the QuanLynx 4.1 software (Waters Corporation, Milford, MA, USA) provided with the instrument. Calibration was performed through linear regression with reciprocal fit weighting to ensure maximum accuracy at the lower concentration range.
|
other
| 98.4 |
The LOD and the LLOQ were determined as the lowest concentration steroid spiked into Schirmer strip calibrators, giving a minimum signal-to-noise ratio (S/N) >3:1 for LOD and >8:1 for LLOQ in replicate analyses (n = 10). The method linearity was evaluated using Schirmer strip calibrators, including two blank Schirmer strip calibrators (with and without addition of internal standard solution) over a period of 5 days. Linearity was defined as deviation from the assigned calibrator values of ≤10%, with the exception of the LLOQ/lowest calibrator for which deviation of ≤15% was accepted. Precision and accuracy for each steroid were evaluated intra- and inter-day using QC material described above. For intra-assay imprecision, a sequence of five replicates of QC in Schirmer strips were analysed in a single batch, and for inter-assay studies, each QC was measured in singleton in batches over a 5-day period . The intra- and inter-assay imprecision were expressed as a CV and deviation from assigned values. The analyte recovery indicates the efficiency of the extraction process and was evaluated following the strategy reported by Matuszewsky et al. . Thus recovery was assessed spanning the analytical range of the assay by analysis of the Schirmer strip QCs and conducted for each steroid. Recovery was calculated as the analyte peak area of the extracted QC in Schirmer strip as a percentage of the peak area of a blank Schirmer strip spiked to the equivalent concentration after the extraction process . Analyses were conducted in triplicate.
|
study
| 100.0 |
Data obtained by LC-MS/MS analysis of tear samples on Schirmer strip from dry eye and control patients were statistically investigated using GraphPad Prism (GraphPad Software, Inc., La Jolla, CA, USA) and and MetaboAnalyst statistical analysis module .
|
study
| 100.0 |
In conclusion, we present for the first time a robust, specific, and selective LC-MS/MS method for the simultaneous quantification of seven steroids from tear samples. The method was, also, applied to a small casuistry of tear samples from 14 female DED patients and 13 female controls, showing a significant reduction in tear response levels for CORT, ADIONE, and 17-OHP in DED patients respect to healthy people. After statistical investigation, our proposal model demonstrated a good predictive power in identifying DED patients. Our data confirm the idea that tears can be used as a precious source of information for the study pathophysiologic state of the eye and of other different disorders.
|
study
| 99.94 |
Influenza A virus undergoes yearly antigenic drift that affects seasonal vaccine effectiveness. Nucleotide sequence changes primarily in the hemagglutinin (HA) gene underlie these yearly changes, and are affected by population-wide immunologic selection [1–5]. Mutation rates have been estimated to occur with a frequency of 5.72 × 10−3 nucleotide substitutions per site per year not only in the HA gene, but also independently in the matrix gene with a similar frequency of 5.39 × 10−3 nucleotide substitutions per site per year [6, 7]. Although molecular methods to detect Influenza virus RNA have targeted conserved areas of the HA and matrix genes, these methods are inherently subject to decreased sensitivity over time, as mutations accumulate in the target sequences of the assay. Examples of reduced sensitivity in molecular assays have been reported due to mutations in the Influenza matrix gene in 2012 and 2013 by Yang et al. in Taiwan , as well as in molecular assays for M. tuberculosis, and enterovirus, Respiratory Syncytial Virus, Hepatitis B, and influenza viruses . In this study, we present data suggesting that sequence differences in the Influenza matrix gene between strains of H3N2 and H1N1 account for differences in the GenMark Respiratory Viral Panel (RVP) matrix gene assay nanoamperes (nAMPs), and that these differences are reflected in the population average mean ± SDs of all positive matrix gene tests over the course of an entire season.
|
study
| 100.0 |
Our objective was to determine whether averaging the semi-quantitative nanoamperes (nAMPS) obtained from the GenMark (Carlsbad, CA) Respiratory Viral Panel (RVP) for all positive results for an entire Influenza season would show meaningful differences due to Influenza sequence drift.
|
study
| 99.94 |
De-identified nAMP data from all patients who had a respiratory virus panel (RVP) (GenMark Diagnostics, Inc. eSensor, Carlsbad, CA) performed between 2012 – February, 2015 at UFHealth Shands Hospital, Gainesville, FL, BayCare Health System, Clearwater, FL and Pathology Consultants of South Broward, Hollywood, FL. All patient sample types (nasopharyngeal (NP) swabs, nasal swabs, throat swabs, sputums, endotracheal suction, bronchoalveolar lavage (BAL), etc.) were included in the study since the intent was to use population averages as a means of controlling for differences in clinical and sample collection variables.
|
study
| 100.0 |
At UFHealth Shands Hospital, 200 μl of patient sample in viral transport medium was extracted with the MagnaPure compact (Roche Diagnostics, Indianapolis, IN), and was eluted in 50 μl of which 5 μl was added to the GenMark RVP assay. At Memorial Healthcare System, respiratory virus samples were extracted utilizing the easyMag (bioMerieux, Durham, NC). Two-hundred microliters of sample (nasopharyngeal swab) collected and transported in viral transport medium was prepared using the on-board protocol and extracted as per manufacturer recommendations. Nasopharyngeal swab samples not placed in viral transport medium were occasionally collected in Liquid Stuarts media. These samples were processed in 2 ml of lysis buffer (bioMerieux, Durham, NC) and prepared in an off-board extraction protocol as per manufacturer recommendations before loading onto the easyMag extractor. Nucleic acid was eluted in 60 ul with 5 ul being used in the GenMark RVP assay. At BayCare Health System 200 μl of patient sample was extracted using the QIAGEN QIAamp MinElute Virus Spin Kit on the semi-automated QiaCube system with on-board lysis. Nucleic acid was eluted in 60 μl and 5 μl were used in the GenMark RVP assay.
|
study
| 99.94 |
The GenMark RVP panel was used according to instructions from the manufacturer. The extracted nucleic acid is reverse transcribed and amplified using viral specific primers with RT-PCR enzyme mix. The amplified DNA is converted to single-stranded DNA via exonuclease digestion and is then combined with a signal buffer containing ferrocene-labeled signal probes that are specific for the different viral targets . A visual description is available at https://www.genmarkdx.com/solutions/technology/esensor/. Amplicon detection is measured by the peak height of a current flowing between the gold electrode and the ferrocene labeled probe, which is brought into proximity with the electrode by the capture probe binding to the target amplicon. Thus the peak current flow in nanoAMPs is a function of the number of targets bound to the capture probes and the tightness of this bond. The technology is sufficiently sensitive to be able to detect single basepair(bp) mutations in cystic fibrosis (https://genmarkdx.com/solutions/panels/xt-8-panels/cystic-fibrosis-genotyping-test/) and in thrombophilia (https://genmarkdx.com/solutions/panels/xt-8-panels/thrombophilia-risk-test/).
|
study
| 99.94 |
We developed a TaqMan PCR for conserved regions of the matrix gene that had no sequence variations in the primers and probe when matched against an environmental H1N1 strain from 2013 to 2014 season isolated at the University of Florida by one of the authors (JL) (Influenza A virus (A/environment/Gainesville/01/2014(H1N1) KJ 195790), and against sequences of A/New York/39/2012 (H3N2) FR-1307, A/Texas/50/2012 (H3N2) FR-1210, and A/Switzerland/9715293/2013 (H3N2) FR-1368 obtained from IRR/ATCC. The sequences for the assay are as follows:Forward primer 89-S: CCG AGA TCG CGC AGA GAC.Reverse primer 239-AS: GCT CAC TGG GCA CGG TG.Probe 177-AS-Pr: ATT GGT CTT GTC TTT AGC CAT TCC ATG AGA G.
|
study
| 99.94 |
The TaqMan assay was used to match viral RNA copy number for supernatant tissue culture fluid for these 4 strains grown in tissue culture. Tissue culture fluids were extracted, adjusted appropriately for differences in concentration as measured by the TaqMan assay and run in the RVP in quadruplicate.
|
study
| 99.94 |
The matrix gene for the 4 Influenza strains were sequenced by traditional Sanger sequencing in the UF Biotechnology Core laboratory facility, using the following primers for an approximately 800 base pair (bp) product:Forward primer 89‐S:CCG AGA TCG CGC AGA GACReverse sequencing primer ATATTC TTCCCT CAT RGA CTC AG
|
study
| 99.9 |
Since the RVP Influenza assay sequences are proprietary, sequences were sent to GenMark Diagnostics who provided data showing the number and location of mismatched base pairs in the Matrix and HA gene sequence in the RVP forward and reverse primers, capture probe and signal probe.
|
other
| 99.9 |
The following Influenza HA gene sequences were obtained from GenBank and were sent to GenMark for matching with their subtype assay sequences: A/Texas/50/2012 (H3N2) KJ942616.1, A/Florida/15/2014(H3N2) KM064336.1, A/Florida/32/2014(H3N2) KR057571.1, A/Florida/21/2014(H3N2) KM972893.1, A/Florida/35/2014(H3N2)KT836943.1 and A/environment/Gainesville/01/2014(H1N1) KJ195788.
|
study
| 99.8 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.