Biology subjects

Narasimhan, G.

Publications and source records attributed to Narasimhan, G..

18 recordsLinked to original sources

bioRxiv2025

PathPCNet: Pathway Principal Component-Based Interpretable Framework for Drug Sensitivity Prediction

Precision medicine aims to identify significant biomarkers and effective drugs tailored to individual genomic profiles, thereby enabling personalized treatment strategies. Drug efficacy is often attributed to drug response, commonly measured as the concentration of a drug required to inhibit a biological activity. In contrast, drug sensitivity reflects how strongly a tumor responds to a drug, where a lower effective dose indicates higher sensitivity. Machine learning-based drug response prediction has the potential to accelerate biomarker discovery and facilitate the development of more effective therapeutics. In this study, we present PathPCNet, a novel interpretable deep learning framework that integrates multi-omics data (copy number variation, mutation, and RNA sequencing) fused with biological pathways, drug molecular structure, and Principal Component Analysis for drug response prediction. Our model achieves a Pearson correlation coefficient of 0.941 and an R-squared of 0.885, outperforming the existing pathway-based approaches. We employ SHAP-based model interpretation to quantify the contributions of omics and drug features, uncovering key pathways and gene-drug interactions associated with resistance mechanisms. These results demonstrate the utility of integrative deep learning models not only for accurate prediction but also for generating biologically meaningful insights, which can advance drug discovery and precision oncology. In addition, the framework also facilitates the identification of important pathways, genes, and atomic attributes of drugs related to drug sensitivity and different cancer types.

bioinformatics↗

bioRxiv2023

A Comprehensive Survey of Scoring Functions for Protein Docking Models

The protein-protein docking problem is fundamental to our understanding of how proteins interact. It is also important because of its relevance to drug discovery and vaccine design. Docking methods typically involve two steps. The first step generates many possible candidate conformations, while the second step evaluates the candidates using some scoring function. Despite numerous experiments with many innovative scoring functions, a good scoring function for docking remains elusive. Deep learning models offer alternatives to using explicit empirical or mathematical functions for scoring protein-protein complexes. In this study, we perform a comprehensive survey of the state-of-the-art scoring functions by considering the most popular and highly performant approaches for scoring protein-protein complexes, evaluating their strengths and weaknesses to aid researchers in understanding progress made in this field. Our survey compares the state-of-the-art in classical and deep learning-based scoring methods.

molecular biology↗

bioRxiv2023

Unfolding and De-confounding: Biologically meaningful causal inference from longitudinal multi-omic networks using METALICA

A key challenge in the analysis of microbiome data is the integration of multi-omic datasets and the discovery of interactions between microbial taxa, their expressed genes, and the metabolites they consume and/or produce. In an effort to improve the state-of-the-art in inferring biologically meaningful multi-omic interactions, we sought to address some of the most fundamental issues in causal inference from longitudinal multi-omics microbiome data sets. We developed METALICA, a suite of tools and techniques that can infer interactions between microbiome entities. METALICA introduces novel unrolling and de-confounding techniques used to uncover multi-omic entities that are believed to act as confounders for some of the relationships that may be inferred using standard causal inferencing tools. The results lend support to predictions about biological models and processes by which microbial taxa interact with each other in a microbiome. The unrolling process helps to identify putative intermediaries (genes and/or metabolites) to explain the interactions between microbes; the de-confounding process identifies putative common causes that may lead to spurious relationships to be inferred. METALICA was applied to the networks inferred by existing causal discovery and network inference algorithms applied to a multi-omics data set resulting from a longitudinal study of IBD microbiomes. The most significant unrollings and de-confoundings were manually validated using the existing literature and databases. ImportanceWe have developed a suite of tools and techniques capable of inferring interactions between microbiome entities. METALICAintroduces novel techniques called unrolling and de-confounding that are employed to uncover multi-omic entities considered to be confounders for some of the relationships that may be inferred using standard causal inferencing tools. To evaluate our method, we conducted tests on the Inflammatory Bowel Disease (IBD) dataset from the iHMP longitudinal study, which we pre-processed in accordance with our previous work.

bioinformatics↗

bioRxiv2023

Genetic determinants of virulence and extensive drug resistance in Pseudomonas aeruginosa PPA14 isolated from eggplant rhizosphere

Pseudomonas aeruginosa is one of the Priority Level I critical pathogens that are least sensitive to antibiotics and can cause fatal hospital-acquired infections. This bacterium is predominantly present in the agricultural ecosystem. However, there are very limited studies on health threats associated with P. aeruginosa strains flourishing in edible plants. Previously, we isolated and characterized 18 P. aeruginosa strains from vegetable plants directly harvested from the farms. In the current work, it has been hypothesized that plant-associated P. aeruginosa harbors genetic determinants for virulence and resistance. To test this hypothesis, in vitro resistome profiles of the plant-associated P. aeruginosa strains were assessed based on the Kirby-Bauer disk diffusion method. Hierarchical clustering analysis was done to identify the plant-associated strains that are phenotypically similar to clinical isolates. An eggplant-associated strain, PPA14, that exhibited high virulence and extensive in vitro resistance against eight antibiotic classes was selected for complete genome analyses. The PPA14 genome was sequenced using the Solexa-Illumina and Oxford-Nanopore platforms, assembled, and annotated. The presence of virulence-related and antibiotic resistance (ABR) genes were predicted using the ABRicate tool and validated based on standard reference databases such as VFDB, NCBI AMRFinderPlus, MEGARes, CARD, and ResFinder. IslandViewer4 tool was used to predict the genes acquired through horizontal gene transfer. Additionally, comparative analyses of all the plant-associated and environmental P. aeruginosa genomes characterized so far were done using the Roary tool. The PPA14 genome size was 6.72 Mbp, encoding 6315 open reading frames. The genome harbored 49 ABR genes, including those coding for multiple families of efflux pumps that collectively confer resistance against 11 antibiotic classes. In addition, we detected 225 virulence-related genes, 83 genomic islands, and 235 unique genes in the PPA14 genome. Over 4% of the PPA14 genome is devoted to conferring virulence and extensive drug resistance. Our report highlights the health threat associated with an eggplant-associated P. aeruginosa.

microbiology↗

bioRxiv2023

PIsToN: Evaluating Protein Binding Interfaces with Transformer Networks

The computational studies of protein binding are widely used to investigate fundamental biological processes and facilitate the development of modern drugs, vaccines, and therapeutics. Scoring functions aim to predict complexes that would be formed by the binding of two biomolecules and to assess and rank the strength of the binding at the interface. Despite past efforts, the accurate prediction and scoring of protein binding interfaces remain a challenge. The physics-based methods are computationally intensive and often have to trade accuracy for computational cost. The possible limitations of current machine learning (ML) methods are ineffective data representation, network architectures, and limited training data. Here, we propose a novel approach called PIsToN (evaluating Protein binding Interfaces with Transformer Networks) that aim to distinguish native-like protein complexes from decoys. Each protein interface is transformed into a collection of 2D images (interface maps), where each image corresponds to a geometric or biochemical property in which pixel intensity represents the feature values. Such a data representation provides atomic-level resolution of relevant protein characteristics. To build hybrid machine learning models, additional empirical-based energy terms are computed and provided as inputs to the neural network. The model is trained on thousands of native and computationally-predicted protein complexes that contain challenging examples. The multi-attention transformer network is also endowed with explainability by highlighting the specific features and binding sites that were the most important for the classification decision. The developed PIsToN model significantly outperforms existing state-of-the-art scoring functions on well-known datasets.

bioinformatics↗

bioRxiv2022

Attention Deficit Hyperactivity Disorder and the gut microbiome: An ecological perspective

Attention Deficit Hyperactivity Disorder (ADHD) is an increasingly prevalent neuropsychiatric disorder characterized by hyperactivity, inattention, and impulsivity. Symptoms emerge from underlying deficiencies in neurocircuitry, and recent research has suggested a role played by the gut microbiome. The gut microbiome is a complex ecosystem of interdependent taxa with an exponentially complex web of interactions involving these taxa, plus host gene and reaction pathways, some of which involve neurotransmitters with roles in ADHD neurocircuitry. Studies have analyzed the ADHD gut microbiome using macroscale metrics such as diversity and composition, and have proposed several biomarkers. Few studies have delved into the complex underlying dynamics ultimately responsible for the emergence of such metrics, leaving a largely incomplete, sometimes contradictory, and ultimately inconclusive picture. We aim to help complete this picture by venturing beyond taxa abundances and into taxa relationships (i.e. cooperation and competition), using a publicly available gut microbiome dataset from 30 Control (15 female, 15 male) and 28 ADHD (15 female, 13 male) undergraduate students. We conduct our study in two parts. We first perform the same macroscale analyses prevalent in ADHD gut microbiome literature (diversity, differential, biomarker, and composition) to observe the degree of correspondence, or any new trends. We then estimate two-way ecological relationships by producing Control and ADHD Microbial Co-occurrence Networks (MCNs), using SparCC correlations (p < 0.01). We perform community detection to find clusters of taxa estimated to mutually cooperate along with their centroids, and centrality calculations to estimate taxa most vital to overall gut ecology. We conclude by summarizing our results, and provide conjectures on how they can guide future experiments, some methods for improving our experiments, and general implications for the field.

microbiology↗

bioRxiv2022

EMoMiS: A Pipeline for Epitope-based Molecular Mimicry Search in Protein Structures with Applications to SARS-CoV-2

MotivationEpitope-based molecular mimicry occurs when an antibody cross-reacts with two different antigens due to structural and chemical similarities. Molecular mimicry between proteins from two viruses can lead to beneficial cross-protection when the antibodies produced by exposure to one also react with the other. On the other hand, mimicry between a protein from a pathogen and a human protein can lead to auto-immune disorders if the antibodies resulting from exposure to the virus end up interacting with host proteins. While cross-protection can suggest the possible reuse of vaccines developed for other pathogens, cross-reaction with host proteins may explain side effects. There are no computational tools available to date for a large-scale search of antibody cross-reactivity. ResultsWe present a comprehensive Epitope-based Molecular Mimicry Search (EMoMiS) pipeline for computational molecular mimicry searches. EMoMiS, when applied to the SARS-CoV-2 Spike protein, identified eight examples of molecular mimicry with viral and human proteins. These findings provide possible explanations for (a) differential severity of COVID-19 caused by cross-protection due to prior vaccinations and/or exposure to other viruses, and (b) commonly seen COVID-19 side effects such as thrombocytopenia and thrombophilia. Our findings are supported by previously reported research but need validation with laboratory experiments. The developed pipeline is generic and can be applied to find mimicry for novel pathogens. It has applications in improving vaccine design. AvailabilityThe developed Epitope-based Molecular Mimicry Search Pipeline (EMoMiS) is available from https://biorg.cs.fiu.edu/emomis/. Contactgiri@cs.fiu.edu

bioinformatics↗

bioRxiv2021

Prior exposure to microcystin alters host gut resistome and is associated with dysregulated immune homeostasis in translatable mouse models

The increased propensity of harmful algal blooms (HABs) and exposure from HABs-cyanotoxin causes human toxicity. It has been associated with the progression of several diseases that encompass the liver, kidneys, and immune system. Recently, a strong association of cyano-HAB toxicity with the altered host gut microbiome has been shown. We tested the hypothesis that prior exposure to cyanotoxin microcystin may alter the microbiome and induce microbiome-host-resistome crosstalk. Using both wild-type and humanized mice, we show that the mice exposed to microcystin had an altered microbiome signature that harbored antimicrobial resistance genes. Host resistome phenotypes such as mefA, msrD, mel, ant6, and tet40 increased in diversity and relative abundance following microcystin exposure. Interestingly, the increased abundance of these genes was traced to resistance to common antibiotics such as tetracycline, macrolides, glycopeptide, and aminoglycosides, crucial for modern-day treatment for several diseases. Increased abundance of these genes was positively associated with increased expression of PD1, a T-cell homeostasis marker, and pleiotropic inflammatory cytokine IL-6 with a concomitant negative association with immunosurveillance markers IL7 and TLR2. Microcystin exposure also caused decreased TLR2, TLR4, and REG3G expressions, increased immunosenescence, and higher systemic levels of IL-6 in both wild-type and humanized mice. In conclusion, the results show a first-ever characterization of the host resistome of microcystin exposure and its connection to host immune status and antibiotic resistance. The results may be crucial for understanding the ability of exposed subjects to fight future bacterial infections and the progression of the debilitating disease in hospital settings.

microbiology↗

bioRxiv2021

Molecular mimicry between Spike and human thrombopoietin may induce thrombocytopenia in COVID-19

SARS-CoV-2 causes COVID-19, a disease curiously resulting in varied symptoms and outcomes, ranging from asymptomatic to fatal. Autoimmunity due to cross-reacting antibodies resulting from molecular mimicry between viral antigens and host proteins may provide an explanation. We computationally investigated molecular mimicry between SARS-CoV-2 Spike and known epitopes. We discovered molecular mimicry hotspots in Spike and highlight two examples with tentative autoimmune potential and implications for understanding COVID-19 complications. We show that a TQLPP motif in Spike and thrombopoietin shares similar antibody binding properties. Antibodies cross-reacting with thrombopoietin may induce thrombocytopenia, a condition observed in COVID-19 patients. Another motif, ELDKY, is shared in multiple human proteins such as PRKG1 and tropomyosin. Antibodies cross-reacting with PRKG1 and tropomyosin may cause known COVID-19 complications such as blood-clotting disorders and cardiac disease, respectively. Our findings illuminate COVID-19 pathogenesis and highlight the importance of considering autoimmune potential when developing therapeutic interventions to reduce adverse reactions.

bioinformatics↗

bioRxiv2021

Multi-run Concrete Autoencoder to Identify Prognostic lncRNAs for 12 Cancers

Long non-coding RNA plays a vital role in changing the expression profiles of various target genes that leads to cancer development. Thus, identifying prognostic lncRNAs related to different cancers might help in developing cancer therapy. To discover the critical lncRNAs that can identify the origin of different cancers, we proposed to use the state-of-the-art deep learning algorithm Concrete Autoencoder (CAE) in an unsupervised setting, which efficiently identifies a subset of the most informative features. However, CAE does not identify reproducible features in different runs due to its stochastic nature. We proposed a multi-run CAE (mrCAE) to identify a stable set of features to address this issue. The assumption is that a feature appearing in multiple runs carries more meaningful information about the data under consideration. The genome-wide lncRNA expression profiles of 12 different types of cancers, a total of 4,768 samples available in The Cancer Genome Atlas (TCGA), were analyzed to discover the key lncRNAs. The lncRNAs identified by multiple runs of CAE were added to the final list of key lncRNAs, which are capable of identifying 12 different cancers. Our results showed that mrCAE performs better in feature selection than single-run CAE, standard autoencoder (AE), and other state-of-the-art feature selection techniques. This study discovered a set of top-ranking 128 lncRNAs that could identify the origin of 12 different cancers with an accuracy of 95%. Survival analysis showed that 76 of 128 lncRNAs have the prognostic capability in differentiating high- and low-risk groups of patients in different cancers. The proposed mrCAE outperformed AE, which select latent features and were thought to be the best tools for dimensionality reduction. By selecting actual features instead of pseudo-features, mrCAE can be valuable for precision medicine. The identified prognostic lncRNAs (this work), mRNAs, miRNAs, and DNA methylated genes (future work) can lead to biomarkers and therapies for different cancers.

bioinformatics↗

bioRxiv2021

Mutation-induced Changes in the Receptor-binding Interface of the SARS-CoV-2 Delta Variant B.1.617.2 and Implications for Immune Evasion

While the vaccination efforts against SARS-CoV-2 infections are ongoing worldwide, new genetic variants of the virus are emerging and spreading. Following the initial surges of the Alpha (B.1.1.7) and the Beta (B.1.351) variants, a more infectious Delta variant (B.1.617.2) is now surging, further deepening the health crises caused by the pandemic. The sharp rise in cases attributed to the Delta variant has made it especially disturbing and is a variant of concern. Fortunately, current vaccines offer protection against known variants of concern, including the Delta variant. However, the Delta variant has exhibited some ability to dodge the immune system as it is found that neutralizing antibodies from prior infections or vaccines are less receptive to binding with the Delta spike protein. Here, we investigated the structural changes caused by the mutations in the Delta variants receptor-binding interface and explored the effects on binding with the ACE2 receptor as well as with neutralizing antibodies. We find that the receptor-binding {beta}-loop-{beta} motif adopts an altered but stable conformation causing separation in some of the antibody binding epitopes. Our study shows reduced binding of neutralizing antibodies and provides a possible mechanism for the immune evasion exhibited by the Delta variant.

biophysics↗

bioRxiv2021

Microbiome Maps: Hilbert Curve Visualizations of Metagenomic Profiles

MotivationAbundance profiles from metagenomic sequencing data synthesize information from billions of sequenced reads coming from thousands of microbial genomes. Analyzing and understanding these profiles can be a challenge since the data they represent are complex. Particularly challenging is their visualization, as existing techniques are inadequate when the taxa number is in the thousands. We present a technique, and accompanying software, for the visualization of metagenomic abundance profiles using a space-filling curve that transforms a profile into an interactive 2D image. ResultsWe created JO_SCPLOWASPERC_SCPLOW, an easy to use tool for the visualization and exploration of metagenomic profiles from DNA sequencing data. It orders taxa using a space-filling Hilbert curve, and creates a "Microbiome Map", where each position in the image represents the abundance of a single taxon from a reference collection. JO_SCPLOWASPERC_SCPLOW can order taxa in multiple ways, and the resulting microbiome maps can highlight "hot spots" of microbes that are dominant in taxonomic clades or biological conditions. We use JO_SCPLOWASPERC_SCPLOW to visualize samples from a variety of microbiome studies, and discuss ways in which microbiome maps can be an invaluable tool to visualize spatial, temporal, disease, and differential profiles. Our approach can create detailed microbiome maps involving hundreds of thousands of microbial reference genomes with the potential to unravel latent relationships (taxonomic, spatio-temporal, functional, and other) that could remain hidden using traditional visualization techniques. The maps can also be converted into animated movies that bring to life the dynamicity of microbiomes. AvailabilityJO_SCPLOWASPERC_SCPLOW is freely available at microbiomemaps.org and via biorg.cs.fiu.edu/jasper Contactcvaldes2@unl.edu; giri@fiu.edu Supplementary informationSupplementary materials are available at microbiomemaps.org

bioinformatics↗

bioRxiv2020

Novel Approach for Microbiome Analysis Using Bacterial Replication Rates and Causal Inference with Applications

MotivationMetagenomics sequencing data can be used to compute not just the relative abundance profile, but also the replication rates of every taxon in the microbiome sample. We investigate how the dynamics implied by the replication rates can be used to understand the antibiotic response in microbiomes, given the significant variation in the types of antibiotics and the types of response by different taxa. The analysis is further expanded by factoring in the resistome of the microbiomes, which can be readily profiled from the metagenomic sequence data. The fact that some antibiotics such as {beta}-lactams target replicating cells makes it even more critical to use replication rates to analyze the antibiotic response. ResultsWe introduce a novel approach for metagenomic analysis that integrates microbial community profiling, replication rate calculation, and causal structural learning to analyze the antibiotic response. First, we developed PO_SCPLOWEC_SCPLOWTRO_SCPLOWIC_SCPLOW, which involves efficient cluster computation of bacterial replication rates from metagenomic sequence data. Second, we integrate the abundance profile, replication profile, resistome profile, and environmental variables to perform causality analysis. Finally, we applied the integrated analysis to the data from an infant gut microbiome study. Conclusions from our analysis are as follows: (i) Microbes tend to lower their replication rates in response to {beta}-lactams; (ii) The presence of antibiotic resistance genes combined with the causality analysis strongly suggest that genes fosA5, oqxA, kpnF, arnA, and acrA provides resistance for the taxon K. pneumoniae, allowing it to replicate and dominate the microbiome after the drug ticarcillin-clavulanate was administered; and (iii) Human and donor milk strongly influence the resistome of the infant gut microbiome. AvailabilityThe pipeline for calculation of bacterial replication rates using the high performance computing framework Slurm is available from https://biorg.cs.fiu.edu/petri/.

bioinformatics↗

bioRxiv2020

Causal Inference in Microbiomes Using Intervention Calculus

Inferring causal effects is critically important in biomedical research as it allows us to move from the typical paradigm of associational studies to causal inference, and can impact treatments and therapeutics. Association patterns can be coincidental and may lead to wrong inferences in complex systems. Microbiomes are highly complex, diverse, and dynamic environments. Microbes are key players in health and diseases. Hence knowledge of genuine causal relationships among the entities in a microbiome, and the impact of internal and external factors on microbial abundance and interactions are essential for understanding disease mechanisms and making treatment recommendations. In this paper, we investigate fundamental causal inference techniques to measure the causal effects of various entities in a microbiome. In particular, we show how to use these techniques on microbiome datasets to study the rise and impact of antibiotic-resistance in microbiomes. Our main contributions include the following. We introduce a novel pipeline for microbiome studies, new ideas for experimental design under weaker assumptions, and data augmentation by context embedding. Our pipeline is robust, different from traditional approaches, and able to predict interventional effects without any controlled experiments. Our work shows the advantages of causal inference in identifying potential pathogenic, beneficial, and antibiotic-resistant bacteria. We validate our results using results that were previously published.

bioinformatics↗

bioRxiv2020

Inferring directional relationships in microbial communities using signed Bayesian networks

BackgroundMicrobe-microbe and host-microbe interactions in a microbiome play a vital role in both health and disease. However, the structure of the microbial community and the colonization patterns are highly complex to infer even under controlled wet laboratory conditions. In this study, we investigate what information, if any, can be provided by a Bayesian Network (BN) about a microbial community. Unlike the previously proposed Co-occurrence Networks (CoNs), BNs are based on conditional dependencies and can help in revealing complex associations. ResultsIn this paper, we propose a way of combining a BN and a CoN to construct a signed Bayesian Network (sBN). We report a surprising association between directed edges in signed BNs and known colonization orders. ConclusionsBNs are powerful tools for community analysis and extracting influences and colonization patterns, even though the analysis only uses an abundance matrix with no temporal information. We conclude that directed edges in sBNs when combined with negative correlations are consistent with and strongly suggestive of colonization order.

bioinformatics↗

bioRxiv2019

Dynamic Bayesian networks for integrating multi-omics time-series microbiome data

A key challenge in the analysis of longitudinal microbiome data is the inference of temporal interactions between microbial taxa, their genes, the metabolites they consume and produce, and host genes. To address these challenges we developed a computational pipeline, PALM, that first aligns multi-omics data and then uses dynamic Bayesian networks (DBNs) to reconstruct a unified model. Our approach overcomes differences in sampling and progression rates, utilizes a biologically-inspired multi-omic framework, reduces the large number of entities and parameters in the DBNs, and validates the learned network. Applying PALM to data collected from inflammatory bowel disease patients, we show that it accurately identifies known and novel interactions. Targeted experimental validations further support a number of the predicted novel metabolite-taxa interactions. Source code and data will be freely available after publication under the MIT Open Source license agreement on our GitHub page. IMPORTANCEWhile a number of large consortia are collecting and profiling several different types of microbiome and genomic time series data, very few methods exist for joint modeling of multi-omics data sets. We developed a new computational pipeline, PALM, which uses Dynamic Bayesian Networks (DBNs) and is designed to integrate multi-omics data from longitudinal microbiome studies. When used to integrate sequence, expression, and metabolomics data from microbiome samples along with host expression data, the resulting models identify interactions between taxa, their genes and the metabolites they produce and consume, and their impact on host expression. We tested the models both by using them to predict future changes in microbiome levels, and by comparing the learned interactions to known interactions in the literature. Finally, we performed experimental validations for a few of the predicted interactions to demonstrate the ability of the method to identify novel relationships and their impact.

bioinformatics↗

bioRxiv2018

Dynamic interaction network inference from longitudinal microbiome data

BackgroundSeveral studies have focused on the microbiota living in environmental niches including human body sites. In many of these studies researchers collect longitudinal data with the goal of understanding not just the composition of the microbiome but also the interactions between the different taxa. However, analysis of such data is challenging and very few methods have been developed to reconstruct dynamic models from time series microbiome data.\n\nResultsHere we present a computational pipeline that enables the integration of data across individuals for the reconstruction of such models. Our pipeline starts by aligning the data collected for all individuals. The aligned profiles are then used to learn a dynamic Bayesian network which represents causal relationships between taxa and clinical variables. Testing our methods on three longitudinal microbiome data sets we show that our pipeline improve upon prior methods developed for this task. We also discuss the biological insights provided by the models which include several known and novel interactions.\n\nConclusionsWe propose a computational pipeline for analyzing longitudinal microbiome data. Our results provide evidence that microbiome alignments coupled with dynamic Bayesian networks improve predictive performance over previous methods and enhance our ability to infer biological relationships within the microbiome and between taxa and clinical factors.

microbiology↗

bioRxiv2017

So you think you can PLS-DA?

BackgroundPartial Least-Squares Discriminant Analysis (PLS-DA) is a popular machine learning tool that is gaining increasing attention as a useful feature selector and classifier. In an effort to understand its strengths and weaknesses, we performed a series of experiments with synthetic data and compared its performance to its close relative from which it was initially invented, namely Principal Component Analysis (PCA). ResultsWe demonstrate that even though PCA ignores the information regarding the class labels of the samples, this unsupervised tool can be remarkably effective as a feature selector. In some cases, it outperforms PLS-DA, which is made aware of the class labels in its input. Our experiments range from looking at the signal-to-noise ratio in the feature selection task, to considering many practical distributions and models encountered when analyzing bioinformatics and clinical data. Other methods were also evaluated. Finally, we analyzed an interesting data set from 396 vaginal microbiome samples where the ground truth for the feature selection was available. All the 3D figures shown in this paper as well as the supplementary ones can be viewed interactively at http://biorg.cs.fiu.edu/plsda ConclusionsOur results highlighted the strengths and weaknesses of PLS-DA in comparison with PCA for different underlying data models.

bioinformatics↗