Biology subjects

Servien, R.

Publications and source records attributed to Servien, R..

4 recordsLinked to original sources

bioRxiv2021

Machine learning models based on molecular descriptors to predict human and environmental toxicological factors in continental freshwater

AO_SCPLOWBSTRACTC_SCPLOWIt is a real challenge for life cycle assessment practitioners to identify all relevant substances contributing to the ecotoxicity. Once this identification has been made, the lack of corresponding ecotoxicity factors can make the results partial and difficult to interpret. So, it is a real and important challenge to provide ecotoxicity factors for a wide range of compounds. Nevertheless, obtaining such factors using experiments is tedious, time-consuming, and made at a high cost. A modeling method that could predict these factors from easy-to-obtain information on each chemical would be of great value. Here, we present such a method, based on machine learning algorithms, that used molecular descriptors to predict two specific endpoints in continental freshwater for ecotoxicological and human impacts. The different tested machine learning algorithms show good performances on a learning database and the non-linear methods tend to outperform the linear ones. The cluster-then-predict approaches usually show the best performances which suggests that these predicted models must be derived for somewhat similar compounds. Finally, predictions were derived from the validated model for compounds with missing toxicity/ecotoxicity factors. HighlightsO_LICharacterization factors (for human health and ecotoxicological impacts) were predicted using molecular descriptors. C_LIO_LISeveral linear or non-linear machine learning methods were compared. C_LIO_LIThe non-linear methods tend to outperform the linear ones using a train and test procedure. Cluster-then-predict approaches often show the best performances, highlighting their usefulness. C_LIO_LIThis methodology was then used to derive characterization factors that were missing for more than a hundred chemicals in USEtox(R). C_LI Graphical abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=107 SRC="FIGDIR/small/453034v6_ufig1.gif" ALT="Figure 1"> View larger version (19K): org.highwire.dtl.DTLVardef@1027386org.highwire.dtl.DTLVardef@1ee9846org.highwire.dtl.DTLVardef@14f2c0aorg.highwire.dtl.DTLVardef@d47a89_HPS_FORMAT_FIGEXP M_FIG C_FIG

biochemistry↗

bioRxiv2020

Joint automatic metabolite identification and quantification of a set of 1H NMR spectra

Metabolomics is a promising approach to characterize phenotypes or to identify biomarkers. It is also easily accessible through NMR, which can provide a comprehensive understanding of the metabolome of any living organisms. However, the analysis of 1H NMR spectrum remains difficult, mainly due to the different problems encountered to perform automatic identification and quantification of metabolites in a reproducible way. In addition, methods that perform automatic identification and quantification of metabolites often do it for one given complex mixture spectrum. Hence, when a set of complex mixture spectra coming from the same experiment has to be processed, the approach is simply repeated independently for every spectrum, despite their resemblance. Here, we present a new method that is the first to identify and quantify metabolites by integrating information coming from several complex spectra of the same experiment. The performances of this new method are then evaluated on both simulated and real datasets. The results show an improvement in the metabolite identification and in the accuracy of metabolite quantifications, especially when the concentration is low. This joint procedure is available in version 2.0 of ASICS package.

bioinformatics↗

bioRxiv2020

The maturity in fetal pigs using a multi-fluid metabolomic approach

In mammalian species, the first days after birth are an important period for survival and the rates of mortality before weaning are high. In pigs, the perinatal deaths average 20% of the litter, with important economic and societal consequences. Among the factors influencing piglet survival at birth, the maturity is likely to be one of the most important. Maturity can be defined as the outcome of complex mechanisms of intra-uterine development and maturation occurring during the last month of gestation. Here, we provide new insights on maturity by studying the end of gestation at two different stages (three weeks before term and close to term) in two breeds of pigs that strongly differ in terms of neonatal survival. Since metabolomics is a promising approach for phenotype characterization or biomarker discovery, we provide a complete understanding of the metabolome of the fetuses in late gestation in three fluids (plasma, urine, and amniotic fluid). We found that biological processes related to amino acid and carbohy-drate metabolisms are critical for piglet maturity. We also confirmed some previously described metabolites associated with delayed growth (e.g., proline and myo-inositol). Altogether, our study proposes new routes for a better characterization of piglet maturity at birth.

developmental biology↗

bioRxiv2018

ASICS: an R package for a whole analysis workflow of 1D 1H NMR spectra

In metabolomics, the detection of new biomarkers from NMR spectra is a promising approach. However, this analysis remains difficult due to the lack of a whole workflow that handles spectra pre-processing, automatic identification and quantification of metabolites and statistical analyses.\n\nWe present ASICS, an R package that contains a complete workflow to analyse spectra from NMR experiments. It contains an automatic approach to identify and quantify metabolites in a complex mixture spectrum and uses the results of the quantification in untargeted and targeted statistical analyses. ASICS was shown to improve the precision of quantification in comparison to existing methods on two independant datasets. In addition, ASICS successfully recovered most metabolites that were found important to explain a two level condition describing the samples by a manual and expert analysis based on bucketting. It also found new relevant metabolites involved in metabolic pathways related to risk factors associated with the conditions.\n\nThis workflow is available through the R package ASICS, available on the Bioconductor platform.

systems biology↗