Comparison of classification methods that combine clinical data and high-dimensional mass spectrometry data

Background

The identification of new diagnostic or prognostic biomarkers is one of the main aims of clinical cancer research. Technologies like mass spectrometry are commonly being used in proteomic research. Mass spectrometry signals show the proteomic profiles of the individuals under study at a given time. These profiles correspond to the recording of a large number of proteins, much larger than the number of individuals. These variables come in addition to or to complete classical clinical variables. The objective of this study is to evaluate and compare the predictive ability of new and existing models combining mass spectrometry data and classical clinical variables. This study was conducted in the context of binary prediction.

Results

To achieve this goal, simulated data as well as a real dataset dedicated to the selection of proteomic markers of steatosis were used to evaluate the methods. The proposed methods meet the challenge of high-dimensional data and the selection of predictive markers by using penalization methods (Ridge, Lasso) and dimension reduction techniques (PLS), as well as a combination of both strategies through sparse PLS in the context of a binary class prediction. The methods were compared in terms of mean classification rate and their ability to select the true predictive values. These comparisons were done on clinical-only models, mass-spectrometry-only models and combined models.

Conclusions

It was shown that models which combine both types of data can be more efficient than models that use only clinical or mass spectrometry data when the sample size of the dataset is large enough.     

Lentivirus‐mediated bifunctional cell labeling for in vivo melanoma study

Lentiviral vectors (LVs) are capable of labeling a broad spectrum of cell types, achieving stable expression of transgenes. However, for in vivo studies, the duration of marker gene expression has been highly variable. We have developed a series of LVs harboring different promoters for expressing reporter gene in mouse cells. Long‐term culture and colony formation of several LV‐labeled mouse melanoma cells showed that promoters derived from mammalian house‐keeping genes, especially those encoding RNA polymerase II (Pol2) and ferritin (FerH), provided the highest consistency for reporter expression. For in vivo studies, primary B16BL6 mouse melanoma were infected with LVs whose luciferase–green fluorescence protein fusion gene (Luc/GFP) was driven by either Pol2 or FerH promoters. When transplanted into syngeneic C57BL/6 mice, Luc/GFP‐labeled B16BL6 mouse melanoma cells can be monitored by bioluminescence imaging in vivo, and GFP‐positive cells can be isolated from the tumors by fluorescence‐activated cell sorter. Pol2‐Luc/GFP labeling, while lower in activity, was more sustainable than FerH‐Luc/GFP labeling in B16BL6 over consecutive passages into mice. We conclude that Pol‐2‐Luc/GFP labeling allows long‐term in vivo monitoring and tumor cell isolation in immunocompetent mouse melanoma models.     

Performance of eDNA assays to detect and quantify an elusive benthic fish in upland streams

The sensitivity and specificity of eDNA-based monitoring, coupled with its potential utility to estimate population density or biomass, makes it a useful tool in invasive species management. In this study, we investigated the potential of the eDNA method to improve the detection of the elusive invasive fish, oriental weatherloach (Misgurnus anguillicaudatus), in a river system where a density gradient of the species occurs. We compared detection rates between eDNA and conventional monitoring methods and examined the relationship between eDNA and abundance in a flowing environment. The eDNA method had a higher site detection rate than conventional methods (63 vs. 38%). Weatherloach eDNA was detected at all sites where the fish has been previously caught and none of the sites where the species has not been caught for the past 7 years. There was an increasing density trend going downstream based on long-term conventional monitoring, but the eDNA concentration in water samples reflected this trend only in a continuous section of the river where impoundments were absent. We did not find a positive relationship between eDNA concentration and contemporary abundance estimates in our study area. A high eDNA concentration was recorded at a site (DVC) which was designated a low density site based on long-term catch data. This discrepancy was a likely result of physical habitat characteristics which influenced the efficiency of the conventional methods used. This study highlighted the challenges of inferring density from eDNA data in flowing water because habitat features may confound results, necessitating careful consideration for results to be useful to management.     

Extended dissimilarity: a method of robust estimation of ecological distances from high beta diversity data

It is widely accepted that reliable ordination of ecological data requires a strong linear or ordinal relationship between the dissimilarity of sites, based on species composition, and the ecological distance between them. Certain dissimilarity measures, having the property that they take a fixed maximum value when sites have no species in common, have been shown to be strongly correlated with ecological distance. For ecological gradients of moderate length (moderate beta diversity), such measures, in conjunction with non-metric multidimensional scaling, will reliably yield successful ordinations. However, as beta diversity increases, more sites have no species in common, and such measures invariably under-estimate ecological distance for such sites. Thus ordinations of data with high species turnover (high beta diversity) may fail.Extended dissimilarities are defined using an iterative adaptation of flexible shortest path adjustment applied to the matrix of dissimilarities with fixed maximum values. By means of theoretical argument and simulations, this is shown to lead to far stronger correlations between the adjusted site dissimilarity and ecological distance for ecological gradients of greater length than previously considered. Hence ordinations of extended dissimilarities, by means of either metric or non-metric scaling techniques, are shown to outperform corresponding ordinations of unadjusted dissimilarities, with the difference increasing with increasing beta diversity.     

Developing Conceptual Models for Complex Ecological Risk Assessments

Conceptual models are representations of the assumed relationship between sources and effects. They serve three important purposes. (1) Their creation compels assessors to think through and clarify their assumptions concerning the activities being assessed. (2) They serve as a communication tool for conveying those assumptions to risk managers and stakeholders. (3) They provide a basis for organizing and conducting the risk assessment. Conceptual models for risk assessments of chemical contaminants have not been problematical because they simply portray the flow of the contaminant from a source to a receptor that experiences toxic effects. However, ecological risk assessments must increasingly deal with direct and indirect effects on multiple endpoint receptors and with multiple complex activities including both physical effects and toxic effects. This complexity may result in conceptual models that are incomplete, ambiguous, or simply too complex to be understood. This paper presents a strategy for creating conceptual models for complex ecological risk assessments that are not only complete but also clear and comprehensible. The strategy also promotes efficiency by creating modular component models for activities, sites, and receptors that can be reused in different combinations in different assessments.     

Genetic diversity of the endangered endemic Agave victoriae-reginae (Agavaceae) in the Chihuahuan Desert

Long-lived perennials are a species-rich, ecologically important component of the North American deserts, yet we know little about their genetic structure, information important for their conservation. Agave victoriae-reginae is an endemic of the Chihuahuan Desert of northern Mexico that is endangered by collection for the ornamental trade. We examined levels and patterns of variation at ten polymorphic allozyme loci in ten populations representing the range of the species. Levels of genetic variation (mean H_e= 0.335) and differentiation (mean F_ST = 0.236) were high. Phenetic clustering suggested the existence of at least three distinct groups of populations. If this pattern of variation is representative of other long-lived desert perennials, it may explain the species richness of this group and will pose a real challenge to gene conservation efforts.     

Hemodynamic Changes in Response to Hyperacute Spinal Trauma in a Swine Model

Acute spinal cord injury (ASCI) is a devastating event that can have severe hemodynamic consequences, depending on location and severity of the lesion. Knowledge of hyperacute hemodynamic changes is important for researchers using porcine models of thoracic ASCI. The goal of this study was to determine the hyperacute hemodynamic changes observed after ASCI when using pigs as their own controls. Five Yucatan gilts were anesthetized, and a dorsal laminectomy performed at T10-T12. Standardized blunt trauma was applied for 5 consecutive min, and hemodynamic variables were collected 5 min before ASCI, and at 2, 4, 6, 8, 10, 20, 30, 60, 80 and 120 min after ASCI. Arterial blood gas samples were collected at 60 min and 10 min before, and at 30 min and between 120 and 240 min after ASCI. Parametric data were analyzed using a mixed effects model with time point as the fixed factor and subject as the random factor. We found no effect on heart rate, pulse pressure, SpO₂, EtCO₂, and respiratory rate between baseline and timepoints after ASCI. Diastolic arterial pressure, mean arterial pressure, and systolic arterial pressure fell significantly by 18%, 16%, and 15%, respectively, at 2 min after ASCI. However, none of the decrements in arterial pressures resulted in hypotension at any time point. Heart rate did not change significantly after ASCI. Blood glucose progressively increased to 50% above baseline between 120 and 240 minutes after ASCI. Low thoracic ASCI caused a consistent and statistically significant but clinically minor hyperacute decrease in arterial pressures (-15%) that did not produce hypotension or metabolic changes suggestive of tissue hypoperfusion. Our findings using this model suggest that mean arterial pressures should be maintained above 85 mm Hg prior to spinal trauma in order to avoid hypotensive states after ASCI.Abbreviations and Acronyms: ASCI, acute spinal cord injury; BE, base excess; BT, core body temperature; CRI, continuous rate infusion; DAP, diastolic arterial pressure; EtCO2, expired partial pressure of carbon dioxide; EtISO, end tidal isoflurane concentration; FLK, fentanyl lidocaine ketamine; Glu, glucose; HR, heart rate; Hb, hemoglobin; Lac, lactate; LRS, lactated Ringer’s solution; MAP, mean arterial pressure; MLK, morphine lidocaine ketamine; PaCO2, arterial partial pressure of carbon dioxide; PaO2, arterial partial pressure of oxygen; PP, pulse pressure; RR, respiratory rate; SAP, systolic arterial pressure; SCI, spinal cord injury     

In the footsteps of sea stars: deciphering the catalogue of proteins involved in underwater temporary adhesion

Sea stars adhere strongly but temporarily to underwater substrata via the secretion of a blend of proteins, forming an adhesive footprint that they leave on the surface after detachment. Their tube feet enclose a duo-gland adhesive system comprising two types of adhesive cells, contributing different layers of the footprint and de-adhesive cells. In this study, we characterized the catalogue of sea star footprint proteins (Sfps) in the species Asterias rubens to gain insights in their potential function. We identified 16 Sfps and mapped their expression to type 1 and/or type 2 adhesive cells or to de-adhesive cells by double fluorescent in situ hybridization. Based on their cellular expression pattern and their conserved functional domains, we propose that the identified Sfps serve different functions during attachment, with two Sfps coupling to the surface, six providing cohesive strength and the rest forming a binding matrix. Immunolabelling of footprints with antibodies directed against one protein of each category confirmed these roles. A de-adhesive gland cell-specific astacin-like proteinase presumably weakens the bond between the adhesive material and the tube foot surface during detachment. Overall, we provide a model for temporary adhesion in sea stars, including a comprehensive list of the proteins involved.