Life thrives in Earth's most inhospitable environments, from boiling hydrothermal vents to hypersaline lakes and frozen polar deserts, thanks to the remarkable adaptations of extremophilic microorganisms. The study of these organisms has rapidly evolved from early cultivation-based discoveries to a data-rich discipline powered by advanced omics technologies. This review comprehensively outlines the current landscape and future directions in extremophile research, emphasizing the pivotal role of bioinformatics, machine learning (ML), and data-driven approaches. We begin by charting the evolution of methodologies, from innovative in situ cultivation techniques and robust biomolecule extraction protocols to modern multi-omics workflows (metagenomics, transcriptomics, proteomics, and metabolomics) that decode the genetic and functional basis of extremophiles. We then catalogue essential bioinformatics resources and specialized databases critical for annotating extremophile genomes and uncovering their unique adaptive strategies, including protein stabilization and syntrophic metabolic relationships. Finally, we explore the transformative potential of artificial intelligence (AI) and ML in overcoming fundamental challenges in the field. These include predicting the functions of uncharacterized "hypothetical" proteins, identifying novel extremozymes, modeling complex genotype-phenotype relationships, and guiding the targeted engineering of industrially relevant strains. By synthesizing insights across these domains, this review highlights how integrating computational biology and AI is poised to unlock the full biotechnological potential of extremophiles and redefine the boundaries of life itself.

Predicting the impact of genomic and drug perturbations in cellular function is crucial for understanding gene functions and drug effects, ultimately leading to improved therapies. To this end, Causal Representation Learning (CRL) constitutes one of the most promising approaches, as it aims to identify the latent factors that causally govern biological systems, thus facilitating the prediction of the effect of unseen perturbations. Yet, current CRL methods fail in reconciling their principled latent representations with known biological processes, leading to models that are not interpretable. To address this major issue, in this work we present SENA-discrepancy-VAE, a model based on the recently proposed CRL method discrepancy-VAE, that produces representations where each latent factor can be interpreted as the (linear) combination of the activity of a (learned) set of biological processes. To this extent, we present an encoder, SENA-$\delta$, that efficiently compute and map biological processes' activity levels to the latent causal factors. We show that SENA-discrepancy-VAE achieves predictive performances on unseen combinations of interventions that are comparable with its original, non-interpretable counterpart, while inferring causal latent factors that are biologically meaningful.

BACKGROUND: Gene-Set Analysis (GSA) is commonly used to analyze high-throughput experiments. However, GSA cannot readily disentangle clusters or pathways due to redundancies in upstream knowledge bases, which hinders comprehensive exploration and interpretation of biological findings. To address this challenge, we developed GeneSetCluster, an R package designed to summarize and integrate GSA results. Over time, we and users as well identified limitations in the original version, such as difficulties in managing redundancies across multiple gene-sets, large computational times, and its lack of accessibility for users without programming expertise. RESULTS: We present GeneSetCluster 2.0, a comprehensive upgrade that delivers methodological, computational, interpretative, and user-experience enhancements. Methodologically, GeneSetCluster 2.0 introduces a novel approach to address duplicated gene-sets and implements a seriation-based clustering algorithm that reorders results, aiding pattern identification. Computationally, the package is optimized for parallel processing, significantly reducing execution time. GeneSetCluster 2.0 enhances cluster annotations by associating clusters with relevant tissues and biological processes to improve biological interpretation, particularly for human and mouse data. To broaden accessibility, we have developed a user-friendly web application enabling non-programmers to use it. This version also ensures seamless integration between the R package, catering to users with programming expertise, and the web application for broader audiences. We evaluated the updates in a single-cell RNA public dataset. CONCLUSION: GeneSetCluster 2.0 offers substantial improvements over its predecessor. Furthermore, by bridging the gap between bioinformaticians and clinicians in multidisciplinary teams, GeneSetCluster 2.0 facilitates collaborative research. The R package and web application, along with detailed installation and usage guides, are available on GitHub ( https://github.com/TranslationalBioinformaticsUnit/GeneSetCluster2.0 ), and the web application can be accessed at https://translationalbio.shinyapps.io/genesetcluster/ .

In the era of precision medicine, it is necessary to understand heterogeneity among patients with complex diseases to improve personalized prevention and management strategies. Here, we introduce ClustAll, a Bioconductor package designed for unsupervised patient stratification using clinical data. ClustAll is based on the previously validated methodology ClustAll, a clustering framework that effectively handles intricacies in clinical data, including mixed data types, missing values, and collinearity. Additionally, ClustAll stands out in its ability to identify multiple patient stratifications within the same population while ensuring their robustness. The updated implementation of ClustAll features S4 classes, parallel computing for enhanced computational efficiency, and user-friendly tools for exploring and comparing stratifications against clinical phenotypes. The performance of ClustAll has been validated using two public clinical datasets, confirming its effectiveness in patient stratification and highlighting its potential impact on clinical management. In summary, ClustAll is a powerful tool for patient stratification in personalized medicine.

MOTIVATION: Genome-wide association studies (GWAS) present several computational and statistical challenges for their data analysis, including knowledge discovery, interpretability, and translation to clinical practice. RESULTS: We develop, apply, and comparatively evaluate an automated machine learning (AutoML) approach, customized for genomic data that delivers reliable predictive and diagnostic models, the set of genetic variants that are important for predictions (called a biosignature), and an estimate of the out-of-sample predictive power. This AutoML approach discovers variants with higher predictive performance compared to standard GWAS methods, computes an individual risk prediction score, generalizes to new, unseen data, is shown to better differentiate causal variants from other highly correlated variants, and enhances knowledge discovery and interpretability by reporting multiple equivalent biosignatures. AVAILABILITY AND IMPLEMENTATION: Code for this study is available at: https://github.com/mensxmachina/autoML-GWAS. JADBio offers a free version at: https://jadbio.com/sign-up/. SNP data can be downloaded from the EGA repository (https://ega-archive.org/). PRS data are found at: https://www.aicrowd.com/challenges/opensnp-height-prediction. Simulation data to study population structure can be found at: https://easygwas.ethz.ch/data/public/dataset/view/1/.

Abstract Molecular gene-expression datasets consist of samples with tens of thousands of measured quantities (i.e., high dimensional data). However, lower-dimensional representations that retain the useful biological information do exist. We present a novel algorithm for such dimensionality reduction called Pathway Activity Score Learning (PASL). The major novelty of PASL is that the constructed features directly correspond to known molecular pathways (genesets in general) and can be interpreted as pathway activity scores . Hence, unlike PCA and similar methods, PASL’s latent space has a fairly straightforward biological interpretation. PASL is shown to outperform in predictive performance the state-of-the-art method (PLIER) on two collections of breast cancer and leukemia gene expression datasets. PASL is also trained on a large corpus of 50000 gene expression samples to construct a universal dictionary of features across different tissues and pathologies. The dictionary validated on 35643 held-out samples for reconstruction error. It is then applied on 165 held-out datasets spanning a diverse range of diseases. The AutoML tool JADBio is employed to show that the predictive information in the PASL-created feature space is retained after the transformation. The code is available at https://github.com/mensxmachina/PASL .

High-throughput technologies such as microarrays and RNA-sequencing (RNA-seq) allow to precisely quantify transcriptomic profiles, generating datasets that are inevitably high-dimensional. In this work, we investigate whether the whole human transcriptome can be represented in a compressed, low dimensional latent space without loosing relevant information. We thus constructed low-dimensional latent feature spaces of the human genome, by utilizing three dimensionality reduction approaches and a diverse set of curated datasets. We applied standard Principal Component Analysis (PCA), kernel PCA and Autoencoder Neural Networks on 1360 datasets from four different measurement technologies. The latent feature spaces are tested for their ability to (a) reconstruct the original data and (b) improve predictive performance on validation datasets not used during the creation of the feature space. While linear techniques show better reconstruction performance, nonlinear approaches, particularly, neural-based models seem to be able to capture non-additive interaction effects, and thus enjoy stronger predictive capabilities. Despite the limited sample size of each dataset and the biological / technological heterogeneity across studies, our results show that low dimensional representations of the human transcriptome can be achieved by integrating hundreds of datasets. The created space is two to three orders of magnitude smaller compared to the raw data, offering the ability of capturing a large portion of the original data variability and eventually reducing computational time for downstream analyses.

BACKGROUND: Feature selection is commonly employed for identifying collectively-predictive biomarkers and biosignatures; it facilitates the construction of small statistical models that are easier to verify, visualize, and comprehend while providing insight to the human expert. In this work we extend established constrained-based, feature-selection methods to high-dimensional "omics" temporal data, where the number of measurements is orders of magnitude larger than the sample size. The extension required the development of conditional independence tests for temporal and/or static variables conditioned on a set of temporal variables. RESULTS: The algorithm is able to return multiple, equivalent solution subsets of variables, scale to tens of thousands of features, and outperform or be on par with existing methods depending on the analysis task specifics. CONCLUSIONS: The use of this algorithm is suggested for variable selection with high-dimensional temporal data.

BACKGROUND: We address the problem of integratively analyzing multiple gene expression, microarray datasets in order to reconstruct gene-gene interaction networks. Integrating multiple datasets is generally believed to provide increased statistical power and to lead to a better characterization of the system under study. However, the presence of systematic variation across different studies makes network reverse-engineering tasks particularly challenging. We contrast two approaches that have been frequently used in the literature for addressing systematic biases: meta-analysis methods, which first calculate opportune statistics on single datasets and successively summarize them, and data-merging methods, which directly analyze the pooled data after removing eventual biases. This comparative evaluation is performed on both synthetic and real data, the latter consisting of two manually curated microarray compendia comprising several E. coli and Yeast studies, respectively. Furthermore, the reconstruction of the regulatory network of the transcription factor Ikaros in human Peripheral Blood Mononuclear Cells (PBMCs) is presented as a case-study. RESULTS: The meta-analysis and data-merging methods included in our experimentations provided comparable performances on both synthetic and real data. Furthermore, both approaches outperformed (a) the naïve solution of merging data together ignoring possible biases, and (b) the results that are expected when only one dataset out of the available ones is analyzed in isolation. Using correlation statistics proved to be more effective than using p-values for correctly ranking candidate interactions. The results from the PBMC case-study indicate that the findings of the present study generalize to different types of network reconstruction algorithms. CONCLUSIONS: Ignoring the systematic variations that differentiate heterogeneous studies can produce results that are statistically indistinguishable from random guessing. Meta-analysis and data merging methods have proved equally effective in addressing this issue, and thus researchers may safely select the approach that best suit their specific application.

MOTIVATION: Variable selection is a typical approach used for molecular-signature and biomarker discovery; however, its application to survival data is often complicated by censored samples. We propose a new algorithm for variable selection suitable for the analysis of high-dimensional, right-censored data called Survival Max-Min Parents and Children (SMMPC). The algorithm is conceptually simple, scalable, based on the theory of Bayesian networks (BNs) and the Markov blanket and extends the corresponding algorithm (MMPC) for classification tasks. The selected variables have a structural interpretation: if T is the survival time (in general the time-to-event), SMMPC returns the variables adjacent to T in the BN representing the data distribution. The selected variables also have a causal interpretation that we discuss. RESULTS: We conduct an extensive empirical analysis of prototypical and state-of-the-art variable selection algorithms for survival data that are applicable to high-dimensional biological data. SMMPC selects on average the smallest variable subsets (less than a dozen per dataset), while statistically significantly outperforming all of the methods in the study returning a manageable number of genes that could be inspected by a human expert. AVAILABILITY: Matlab and R code are freely available from http://www.mensxmachina.org

BACKGROUND: Recent technological advances in DNA sequencing and genotyping have led to the accumulation of a remarkable quantity of data on genetic polymorphisms. However, the development of new statistical and computational tools for effective processing of these data has not been equally as fast. In particular, Machine Learning literature is limited to relatively few papers which are focused on the development and application of data mining methods for the analysis of genetic variability. On the other hand, these papers apply to genetic data procedures which had been developed for a different kind of analysis and do not take into account the peculiarities of population genetics. The aim of our study was to define a new similarity measure, specifically conceived for measuring the similarity between the genetic profiles of two groups of subjects (i.e., cases and controls) taking into account that genetic profiles are usually distributed in a population group according to the Hardy Weinberg equilibrium. RESULTS: We set up a new kernel function consisting of a similarity measure between groups of subjects genotyped for numerous genetic loci. This measure weighs different genetic profiles according to the estimates of gene frequencies at Hardy-Weinberg equilibrium in the population. We named this function the "Hardy-Weinberg kernel". The effectiveness of the Hardy-Weinberg kernel was compared to the performance of the well established linear kernel. We found that the Hardy-Weinberg kernel significantly outperformed the linear kernel in a number of experiments where we used either simulated data or real data. CONCLUSION: The "Hardy-Weinberg kernel" reported here represents one of the first attempts at incorporating genetic knowledge into the definition of a kernel function designed for the analysis of genetic data. We show that the best performance of the "Hardy-Weinberg kernel" is observed when rare genotypes have different frequencies in cases and controls. The ability to capture the effect of rare genotypes on phenotypic traits might be a very important and useful feature, as most of the current statistical tools loose most of their statistical power when rare genotypes are involved in the susceptibility to the trait under study.