A review of variable selection methods in Partial Least Squares Regression

Variable selection on high throughput biological data, such as gene expression or single nucleotide polymorphisms (SNPs), becomes inevitable to select relevant information and, therefore, to better characterize diseases or assess genetic structure. There are different ways to perform variable selection in large data sets. Statistical tests are commonly used to identify differentially expressed features for explanatory purposes, whereas Machine Learning wrapper approaches can be used for predictive purposes. In the case of multiple highly correlated variables, another option is to use multivariate exploratory approaches to give more insight into cell biology, biological pathways or complex traits. A simple extension of a sparse PLS exploratory approach is proposed to perform variable selection in a multiclass classification framework. sPLS-DA has a classification performance similar to other wrapper or sparse discriminant analysis approaches on public microarray and SNP data sets. More importantly, sPLS-DA is clearly competitive in terms of computational efficiency and superior in terms of interpretability of the results via valuable graphical outputs. sPLS-DA is available in the R package mixOmics, which is dedicated to the analysis of large biological data sets.

/pdf/sparse-pls-discriminant-analysis-biologically-relevant-57wqz5nk5k.pdf

Sparse PLS discriminant analysis: biologically relevant feature selection and graphical displays for multiclass problems

SLEP: Sparse Learning with Efficient Projections

The biological carbon pump is the process by which CO2 is transformed to organic carbon via photosynthesis, exported through sinking particles, and finally sequestered in the deep ocean. While the intensity of the pump correlates with plankton community composition, the underlying ecosystem structure driving the process remains largely uncharacterized. Here we use environmental and metagenomic data gathered during the Tara Oceans expedition to improve our understanding of carbon export in the oligotrophic ocean. We show that specific plankton communities, from the surface and deep chlorophyll maximum, correlate with carbon export at 150 m and highlight unexpected taxa such as Radiolaria and alveolate parasites, as well as Synechococcus and their phages, as lineages most strongly associated with carbon export in the subtropical, nutrient-depleted, oligotrophic ocean. Additionally, we show that the relative abundance of a few bacterial and viral genes can predict a significant fraction of the variability in carbon export in these regions.

/pdf/plankton-networks-driving-carbon-export-in-the-oligotrophic-n2jwa3z754.pdf

Plankton networks driving carbon export in the oligotrophic ocean.

Mixed fermentations using controlled inoculation of Saccharomyces cerevisiae starter cultures and non-Saccharomyces yeasts represent a feasible way towards improving the complexity and enhancing the particular and specific characteristics of wines. The profusion of selected starter cultures has allowed the more widespread use of inoculated fermentations, with consequent improvements to the control of the fermentation process, and the use of new biotechnological processes in winemaking. Over the last few years, as a consequence of the re-evaluation of the role of non-Saccharomyces yeasts in winemaking, there have been several studies that have evaluated the use of controlled mixed fermentations using Saccharomyces and different non-Saccharomyces yeast species from the wine environment. The combined use of different species often results in unpredictable compounds and/or different levels of fermentation products being produced, which can affect both the chemical and the aromatic composition of wines. Moreover, possible synergistic interactions between different yeasts might provide a tool for the implementation of new fermentation technologies. Thus, knowledge of the Saccharomyces and non-Saccharomyces wine yeast interactions during wine fermentation needs to be improved. To reach this goal, further investigations into the genetic and physiological background of such non-Saccharomyces wine yeasts are needed, so as to apply '-omics' approaches to mixed culture fermentations.

https://onlinelibrary.wiley.com/doi/pdf/10.1111/j.1567-1364.2009.00579.x

Controlled mixed culture fermentation: a new perspective on the use of non-Saccharomyces yeasts in winemaking.

Recent biotechnology advances allow for multiple types of omics data, such as transcriptomic, proteomic or metabolomic data sets to be integrated. The problem of feature selection has been addressed several times in the context of classification, but needs to be handled in a specific manner when integrating data. In this study, we focus on the integration of two-block data that are measured on the same samples. Our goal is to combine integration and simultaneous variable selection of the two data sets in a one-step procedure using a Partial Least Squares regression (PLS) variant to facilitate the biologists' interpretation. A novel computational methodology called ;;sparse PLS" is introduced for a predictive analysis to deal with these newly arisen problems. The sparsity of our approach is achieved with a Lasso penalization of the PLS loading vectors when computing the Singular Value Decomposition. Sparse PLS is shown to be effective and biologically meaningful. Comparisons with classical PLS are performed on a simulated data set and on real data sets. On one data set, a thorough biological interpretation of the obtained results is provided. We show that sparse PLS provides a valuable variable selection tool for highly dimensional data sets.

/pdf/a-sparse-pls-for-variable-selection-when-integrating-omics-3tb4hxu7tb.pdf

A Sparse PLS for Variable Selection when Integrating Omics Data

'Omics' tools provide novel opportunities for system-wide analysis of complex cellular functions. Secondary metabolism is an example of a complex network of biochemical pathways, which, although well mapped from a biochemical point of view, is not well understood with regards to its physiological roles and genetic and biochemical regulation. Many of the metabolites produced by this network such as higher alcohols and esters are significant aroma impact compounds in fermentation products, and different yeast strains are known to produce highly divergent aroma profiles. Here, we investigated whether we can predict the impact of specific genes of known or unknown function on this metabolic network by combining whole transcriptome and partial exo-metabolome analysis. For this purpose, the gene expression levels of five different industrial wine yeast strains that produce divergent aroma profiles were established at three different time points of alcoholic fermentation in synthetic wine must. A matrix of gene expression data was generated and integrated with the concentrations of volatile aroma compounds measured at the same time points. This relatively unbiased approach to the study of volatile aroma compounds enabled us to identify candidate genes for aroma profile modification. Five of these genes, namely YMR210W, BAT1, AAD10, AAD14 and ACS1 were selected for overexpression in commercial wine yeast, VIN13. Analysis of the data show a statistically significant correlation between the changes in the exo-metabome of the overexpressing strains and the changes that were predicted based on the unbiased alignment of transcriptomic and exo-metabolomic data. The data suggest that a comparative transcriptomics and metabolomics approach can be used to identify the metabolic impacts of the expression of individual genes in complex systems, and the amenability of transcriptomic data to direct applications of biotechnological relevance.

/pdf/linking-gene-regulation-and-the-exo-metabolome-a-comparative-4x5ch3hhbz.pdf

Linking gene regulation and the exo-metabolome: A comparative transcriptomics approach to identify genes that impact on the production of volatile aroma compounds in yeast

Exploring the phenotypic space of non-Saccharomyces wine yeast biodiversity

The conversion of grape sugar to ethanol and carbon dioxide is the primary biochemical reaction in alcoholic wine fermentation, but microbial interactions, as well as complex secondary metabolic reactions, are equally relevant in terms of the composition of the final wine produced. The chemical composition of a wine determines the taste, flavour and aroma of the product, and is determined by many factors such as grape variety, geographical and viticultural conditions, microbial ecology of the grapes and of the fermentation processes, and the winemaking practices. Through the years, major advances have been made in understanding the biochemistry, ecology, physiology and molecular biology of the various yeast strains involved in wine production, and how these yeasts affect wine chemistry and wine sensory properties. However, many important aspects of the impact of yeast on specific wine-relevant sensory parameters remain little understood. One of these areas of limited knowledge is the contribution of individual wine yeast strains to the total organic acid profile of wine. Wine quality is indeed very directly linked to what wine tasters frequently refer to as the sugar–acid balance. The total acidity of a wine is therefore of prime sensory importance, and acidity adjustments are a frequent and legal practice in many wineries. However, the total acidity is the result of the sum of all the individual organic acids that are present in wine. Importantly, each of these acids has its own sensory attributes, with descriptors ranging from fresh to sour to metallic. It is therefore important not only to consider total acidity, but also the contribution of each individual acid to the overall acid profile of the wine. This review will summarise the current knowledge about the origin, synthesis and analysis of organic acids in wine, as well as on the management of wine acidity.

/pdf/organic-acid-metabolism-and-the-impact-of-fermentation-zbiwbzrsug.pdf

Organic Acid Metabolism and the Impact of Fermentation Practices on Wine Acidity - A Review

The impact of co-inoculation with Oenococcus oeni on the trancriptome of Saccharomyces cerevisiae and on the flavour-active metabolite profiles during fermentation in synthetic must

Debra Rossouw

Papers

A Sparse PLS for Variable Selection when Integrating Omics Data

Linking gene regulation and the exo-metabolome: A comparative transcriptomics approach to identify genes that impact on the production of volatile aroma compounds in yeast

Exploring the phenotypic space of non-Saccharomyces wine yeast biodiversity

Organic Acid Metabolism and the Impact of Fermentation Practices on Wine Acidity - A Review

The impact of co-inoculation with Oenococcus oeni on the trancriptome of Saccharomyces cerevisiae and on the flavour-active metabolite profiles during fermentation in synthetic must