Drugst.One - a plug-and-play solution for online systems medicine and network-based drug repurposing.

In recent decades, the development of new drugs has become increasingly expensive and inefficient, and the molecular mechanisms of most pharmaceuticals remain poorly understood. In response, computational systems and network medicine tools have emerged to identify potential drug repurposing candidates. However, these tools often require complex installation and lack intuitive visual network mining capabilities. To tackle these challenges, we introduce Drugst.One, a platform that assists specialized computational medicine tools in becoming user-friendly, web-based utilities for drug repurposing. With just three lines of code, Drugst.One turns any systems biology software into an interactive web tool for modeling and analyzing complex protein-drug-disease networks. Demonstrating its broad adaptability, Drugst.One has been successfully integrated with 21 computational systems medicine tools. Available at https://drugst.one, Drugst.One has significant potential for streamlining the drug discovery process, allowing researchers to focus on essential aspects of pharmaceutical treatment research.

Diurnal rhythmicity of infant fecal microbiota and metabolites: A randomized controlled interventional trial with infant formula.

Microbiota assembly in the infant gut is influenced by diet. Breastfeeding and human breastmilk oligosaccharides promote the colonization of beneficial bifidobacteria. Infant formulas are supplemented with bifidobacteria or complex oligosaccharides, notably galacto-oligosaccharides (GOS), to mimic breast milk. To compare microbiota development across feeding modes, this randomized controlled intervention study (German Clinical Trial DRKS00012313) longitudinally sampled infant stool during the first year of life, revealing similar fecal bacterial communities between formula- and breast-fed infants (N = 210) but differences across age. Infant formula containing GOS sustained high levels of bifidobacteria compared with formula containing B. longum and B. breve or placebo. Metabolite and bacterial profiling revealed 24-h oscillations and circadian networks. Rhythmicity in bacterial diversity, specific taxa, and functional pathways increased with age and was strongest following breastfeeding and GOS supplementation. Circadian rhythms in dominant taxa were further maintained ex vivo in a chemostat model. Hence, microbiota rhythmicity develops early in life and is impacted by diet.

Cracking the black box of deep sequence-based protein-protein interaction prediction.

Identifying protein-protein interactions (PPIs) is crucial for deciphering biological pathways. Numerous prediction methods have been developed as cheap alternatives to biological experiments, reporting surprisingly high accuracy estimates. We systematically investigated how much reproducible deep learning models depend on data leakage, sequence similarities and node degree information, and compared them with basic machine learning models. We found that overlaps between training and test sets resulting from random splitting lead to strongly overestimated performances. In this setting, models learn solely from sequence similarities and node degrees. When data leakage is avoided by minimizing sequence similarities between training and test set, performances become random. Moreover, baseline models directly leveraging sequence similarity and network topology show good performances at a fraction of the computational cost. Thus, we advocate that any improvements should be reported relative to baseline methods in the future. Our findings suggest that predicting PPIs remains an unsolved task for proteins showing little sequence similarity to previously studied proteins, highlighting that further experimental research into the 'dark' protein interactome and better computational methods are needed.

Inference of differential gene regulatory networks using boosted differential trees.

Summary: Diseases can be caused by molecular perturbations that induce specific changes in regulatory interactions and their coordinated expression, also referred to as network rewiring. However, the detection of complex changes in regulatory connections remains a challenging task and would benefit from the development of novel nonparametric approaches. We develop a new ensemble method called BoostDiff (boosted differential regression trees) to infer a differential network discriminating between two conditions. BoostDiff builds an adaptively boosted (AdaBoost) ensemble of differential trees with respect to a target condition. To build the differential trees, we propose differential variance improvement as a novel splitting criterion. Variable importance measures derived from the resulting models are used to reflect changes in gene expression predictability and to build the output differential networks. BoostDiff outperforms existing differential network methods on simulated data evaluated in four different complexity settings. We then demonstrate the power of our approach when applied to real transcriptomics data in COVID-19, Crohn's disease, breast cancer, prostate adenocarcinoma, and stress response in Bacillus subtilis. BoostDiff identifies context-specific networks that are enriched with genes of known disease-relevant pathways and complements standard differential expression analyses. Availability and implementation: BoostDiff is available at https://github.com/scibiome/boostdiff_inference.

Making mouse transcriptomics deconvolution accessible with immunedeconv.

Summary: Transcriptome deconvolution has emerged as a reliable technique to estimate cell-type abundances from bulk RNA sequencing data. Unlike their human equivalents, methods to quantify the cellular composition of complex tissues from murine transcriptomics are sparse and sometimes not easy to use. We extended the immunedeconv R package to facilitate the deconvolution of mouse transcriptomics, enabling the quantification of murine immune-cell types using 13 different methods. Through immunedeconv, we further offer the possibility of tweaking cell signatures used by deconvolution methods, providing custom annotations tailored for specific cell types and tissues. These developments strongly facilitate the study of the immune-cell composition of mouse models and further open new avenues in the investigation of the cellular composition of other tissues and organisms. Availability and implementation: The R package and the documentation are available at https://github.com/omnideconv/immunedeconv.

On the limits of 16S rRNA gene-based metagenome prediction and functional profiling.

Molecular profiling techniques such as metagenomics, metatranscriptomics or metabolomics offer important insights into the functional diversity of the microbiome. In contrast, 16S rRNA gene sequencing, a widespread and cost-effective technique to measure microbial diversity, only allows for indirect estimation of microbial function. To mitigate this, tools such as PICRUSt2, Tax4Fun2, PanFP and MetGEM infer functional profiles from 16S rRNA gene sequencing data using different algorithms. Prior studies have cast doubts on the quality of these predictions, motivating us to systematically evaluate these tools using matched 16S rRNA gene sequencing, metagenomic datasets, and simulated data. Our contribution is threefold: (i) using simulated data, we investigate if technical biases could explain the discordance between inferred and expected results; (ii) considering human cohorts for type two diabetes, colorectal cancer and obesity, we test if health-related differential abundance measures of functional categories are concordant between 16S rRNA gene-inferred and metagenome-derived profiles and; (iii) since 16S rRNA gene copy number is an important confounder in functional profiles inference, we investigate if a customised copy number normalisation with the rrnDB database could improve the results. Our results show that 16S rRNA gene-based functional inference tools generally do not have the necessary sensitivity to delineate health-related functional changes in the microbiome and should thus be used with care. Furthermore, we outline important differences in the individual tools tested and offer recommendations for tool selection.

Blood transcriptomics analysis offers insights into variant-specific immune response to SARS-CoV-2.

Bulk RNA sequencing (RNA-seq) of blood is typically used for gene expression analysis in biomedical research but is still rarely used in clinical practice. In this study, we propose that RNA-seq should be considered a diagnostic tool, as it offers not only insights into aberrant gene expression and splicing but also delivers additional readouts on immune cell type composition as well as B-cell and T-cell receptor (BCR/TCR) repertoires. We demonstrate that RNA-seq offers insights into a patient's immune status via integrative analysis of RNA-seq data from patients infected with various SARS-CoV-2 variants (in total 196 samples with up to 200 million reads sequencing depth). We compare the results of computational cell-type deconvolution methods (e.g., MCP-counter, xCell, EPIC, quanTIseq) to complete blood count data, the current gold standard in clinical practice. We observe varying levels of lymphocyte depletion and significant differences in neutrophil levels between SARS-CoV-2 variants. Additionally, we identify B and T cell receptor (BCR/TCR) sequences using the tools MiXCR and TRUST4 to show that-combined with sequence alignments and BLASTp-they could be used to classify a patient's disease. Finally, we investigated the sequencing depth required for such analyses and concluded that 10 million reads per sample is sufficient. In conclusion, our study reveals that computational cell-type deconvolution and BCR/TCR methods using bulk RNA-seq analyses can supplement missing CBC data and offer insights into immune responses, disease severity, and pathogen-specific immunity, all achievable with a sequencing depth of 10 million reads per sample.

Comprehensive benchmark of differential transcript usage analysis for static and dynamic conditions.

RNA sequencing offers unique insights into transcriptome diversity, and a plethora of tools have been developed to analyze alternative splicing. One important task is to detect changes in the relative transcript abundance in differential transcript usage (DTU) analysis. The choice of the right analysis tool is non-trivial and depends on experimental factors such as the availability of single- or paired-end and bulk or single-cell data. To help users select the most promising tool for their task, we performed a comprehensive benchmark of DTU detection tools. We cover a wide array of experimental settings, using simulated bulk and single-cell RNA-seq data as well as real transcriptomics datasets, including time-series data. Our results suggest that DEXSeq, edgeR, and LimmaDS are better choices for paired-end data, while DSGseq and DEXSeq can be used for single-end data. In single-cell simulation settings, we showed that satuRn performs better than DTUrtle. In addition, we showed that Spycone is optimal for time series DTU/IS analysis based on the evidence provided using GO terms enrichment analysis.

Blood transcriptomics analysis offers insights into variant-specific immune response to SARS-CoV-2.

Bulk RNA sequencing (RNA-seq) of blood is typically used for gene expression analysis in biomedical research but is still rarely used in clinical practice. In this study, we argue that RNA-seq should be considered a routine diagnostic tool, as it offers not only insights into aberrant gene expression and splicing but also delivers additional readouts on immune cell type composition as well as B-cell and T-cell receptor (BCR/TCR) repertoires. We demonstrate that RNA-seq offers vital insights into a patient's immune status via integrative analysis of RNA-seq data from patients infected with various SARS-CoV-2 variants (in total 240 samples with up to 200 million reads sequencing depth). We compare the results of computational cell-type deconvolution methods (e.g., MCP-counter, xCell, EPIC, quanTIseq) to complete blood count data, the current gold standard in clinical practice. We observe varying levels of lymphocyte depletion and significant differences in neutrophil levels between SARS-CoV-2 variants. Additionally, we identify B and T cell receptor (BCR/TCR) sequences using the tools MiXCR and TRUST4 to show that - combined with sequence alignments and pBLAST - they could be used to classify a patient's disease. Finally, we investigated the sequencing depth required for such analyses and concluded that 10 million reads per sample is sufficient. In conclusion, our study reveals that computational cell-type deconvolution and BCR/TCR methods using bulk RNA-seq analyses can supplement missing CBC data and offer insights into immune responses, disease severity, and pathogen-specific immunity, all achievable with a sequencing depth of 10 million reads per sample.

Network medicine-based epistasis detection in complex diseases: ready for quantum computing.

Most heritable diseases are polygenic. To comprehend the underlying genetic architecture, it is crucial to discover the clinically relevant epistatic interactions (EIs) between genomic single nucleotide polymorphisms (SNPs)1-3. Existing statistical computational methods for EI detection are mostly limited to pairs of SNPs due to the combinatorial explosion of higher-order EIs. With NeEDL (network-based epistasis detection via local search), we leverage network medicine to inform the selection of EIs that are an order of magnitude more statistically significant compared to existing tools and consist, on average, of five SNPs. We further show that this computationally demanding task can be substantially accelerated once quantum computing hardware becomes available. We apply NeEDL to eight different diseases and discover genes (affected by EIs of SNPs) that are partly known to affect the disease, additionally, these results are reproducible across independent cohorts. EIs for these eight diseases can be interactively explored in the Epistasis Disease Atlas (https://epistasis-disease-atlas.com). In summary, NeEDL is the first application that demonstrates the potential of seamlessly integrated quantum computing techniques to accelerate biomedical research. Our network medicine approach detects higher-order EIs with unprecedented statistical and biological evidence, yielding unique insights into polygenic diseases and providing a basis for the development of improved risk scores and combination therapies.

Multi-omics regulatory network inference in the presence of missing data.

A key problem in systems biology is the discovery of regulatory mechanisms that drive phenotypic behaviour of complex biological systems in the form of multi-level networks. Modern multi-omics profiling techniques probe these fundamental regulatory networks but are often hampered by experimental restrictions leading to missing data or partially measured omics types for subsets of individuals due to cost restrictions. In such scenarios, in which missing data is present, classical computational approaches to infer regulatory networks are limited. In recent years, approaches have been proposed to infer sparse regression models in the presence of missing information. Nevertheless, these methods have not been adopted for regulatory network inference yet. In this study, we integrated regression-based methods that can handle missingness into KiMONo, a Knowledge guided Multi-Omics Network inference approach, and benchmarked their performance on commonly encountered missing data scenarios in single- and multi-omics studies. Overall, two-step approaches that explicitly handle missingness performed best for a wide range of random- and block-missingness scenarios on imbalanced omics-layers dimensions, while methods implicitly handling missingness performed best on balanced omics-layers dimensions. Our results show that robust multi-omics network inference in the presence of missing data with KiMONo is feasible and thus allows users to leverage available multi-omics data to its full extent.

Alternative splicing impacts microRNA regulation within coding regions.

MicroRNAs (miRNAs) are small non-coding RNA molecules that bind to target sites in different gene regions and regulate post-transcriptional gene expression. Approximately 95% of human multi-exon genes can be spliced alternatively, which enables the production of functionally diverse transcripts and proteins from a single gene. Through alternative splicing, transcripts might lose the exon with the miRNA target site and become unresponsive to miRNA regulation. To check this hypothesis, we studied the role of miRNA target sites in both coding and non-coding regions using six cancer data sets from The Cancer Genome Atlas (TCGA) and Parkinson's disease data from PPMI. First, we predicted miRNA target sites on mRNAs from their sequence using TarPmiR. To check whether alternative splicing interferes with this regulation, we trained linear regression models to predict miRNA expression from transcript expression. Using nested models, we compared the predictive power of transcripts with miRNA target sites in the coding regions to that of transcripts without target sites. Models containing transcripts with target sites perform significantly better. We conclude that alternative splicing does interfere with miRNA regulation by skipping exons with miRNA target sites within the coding region.

Computational tools for inferring transcription factor activity.

Transcription factors (TFs) are essential players in orchestrating the regulatory landscape in cells. Still, their exact modes of action and dependencies on other regulatory aspects remain elusive. Since TFs act cell type-specific and each TF has its own characteristics, untangling their regulatory interactions from an experimental point of view is laborious and convoluted. Thus, there is an ongoing development of computational tools that estimate transcription factor activity (TFA) from a variety of data modalities, either based on a mapping of TFs to their putative target genes or in a genome-wide, gene-unspecific fashion. These tools can help to gain insights into TF regulation and to prioritize candidates for experimental validation. We want to give an overview of available computational tools that estimate TFA, illustrate examples of their application, debate common result validation strategies, and discuss assumptions and concomitant limitations.

circRNA-sponging: a pipeline for extensive analysis of circRNA expression and their role in miRNA sponging.

Motivation: Circular RNAs (circRNAs) are long noncoding RNAs (lncRNAs) often associated with diseases and considered potential biomarkers for diagnosis and treatment. Among other functions, circRNAs have been shown to act as microRNA (miRNA) sponges, preventing the role of miRNAs that repress their targets. However, there is no pipeline to systematically assess the sponging potential of circRNAs. Results: We developed circRNA-sponging, a nextflow pipeline that (i) identifies circRNAs via backsplicing junctions detected in RNA-seq data, (ii) quantifies their expression values in relation to their linear counterparts spliced from the same gene, (iii) performs differential expression analysis, (iv) identifies and quantifies miRNA expression from miRNA-sequencing (miRNA-seq) data, (v) predicts miRNA binding sites on circRNAs, (vi) systematically investigates potential circRNA-miRNA sponging events, (vii) creates a network of competing endogenous RNAs and (viii) identifies potential circRNA biomarkers. We showed the functionality of the circRNA-sponging pipeline using RNA sequencing data from brain tissues, where we identified two distinct types of circRNAs characterized by a specific ratio of the number of the binding site to the length of the transcript. The circRNA-sponging pipeline is the first end-to-end pipeline to identify circRNAs and their sponging systematically with raw total RNA-seq and miRNA-seq files, allowing us to better indicate the functional impact of circRNAs as a routine aspect in transcriptomic research. Availability and implementation: https://github.com/biomedbigdata/circRNA-sponging. Supplementary information: Supplementary data are available at Bioinformatics Advances online.

The FeatureCloud Platform for Federated Learning in Biomedicine: Unified Approach.

BACKGROUND: Machine learning and artificial intelligence have shown promising results in many areas and are driven by the increasing amount of available data. However, these data are often distributed across different institutions and cannot be easily shared owing to strict privacy regulations. Federated learning (FL) allows the training of distributed machine learning models without sharing sensitive data. In addition, the implementation is time-consuming and requires advanced programming skills and complex technical infrastructures. OBJECTIVE: Various tools and frameworks have been developed to simplify the development of FL algorithms and provide the necessary technical infrastructure. Although there are many high-quality frameworks, most focus only on a single application case or method. To our knowledge, there are no generic frameworks, meaning that the existing solutions are restricted to a particular type of algorithm or application field. Furthermore, most of these frameworks provide an application programming interface that needs programming knowledge. There is no collection of ready-to-use FL algorithms that are extendable and allow users (eg, researchers) without programming knowledge to apply FL. A central FL platform for both FL algorithm developers and users does not exist. This study aimed to address this gap and make FL available to everyone by developing FeatureCloud, an all-in-one platform for FL in biomedicine and beyond. METHODS: The FeatureCloud platform consists of 3 main components: a global frontend, a global backend, and a local controller. Our platform uses a Docker to separate the local acting components of the platform from the sensitive data systems. We evaluated our platform using 4 different algorithms on 5 data sets for both accuracy and runtime. RESULTS: FeatureCloud removes the complexity of distributed systems for developers and end users by providing a comprehensive platform for executing multi-institutional FL analyses and implementing FL algorithms. Through its integrated artificial intelligence store, federated algorithms can easily be published and reused by the community. To secure sensitive raw data, FeatureCloud supports privacy-enhancing technologies to secure the shared local models and assures high standards in data privacy to comply with the strict General Data Protection Regulation. Our evaluation shows that applications developed in FeatureCloud can produce highly similar results compared with centralized approaches and scale well for an increasing number of participating sites. CONCLUSIONS: FeatureCloud provides a ready-to-use platform that integrates the development and execution of FL algorithms while reducing the complexity to a minimum and removing the hurdles of federated infrastructure. Thus, we believe that it has the potential to greatly increase the accessibility of privacy-preserving and distributed data analyses in biomedicine and beyond.

Alternative splicing analysis benchmark with DICAST.

Alternative splicing is a major contributor to transcriptome and proteome diversity in health and disease. A plethora of tools have been developed for studying alternative splicing in RNA-seq data. Previous benchmarks focused on isoform quantification and mapping. They neglected event detection tools, which arguably provide the most detailed insights into the alternative splicing process. DICAST offers a modular and extensible framework for analysing alternative splicing integrating eleven splice-aware mapping and eight event detection tools. We benchmark all tools extensively on simulated as well as whole blood RNA-seq data. STAR and HISAT2 demonstrated the best balance between performance and run time. The performance of event detection tools varies widely with no tool outperforming all others. DICAST allows researchers to employ a consensus approach to consider the most successful tools jointly for robust event detection. Furthermore, we propose the first reporting standard to unify existing formats and to guide future tool development.

Drugst.One - A plug-and-play solution for online systems medicine and network-based drug repurposing.

In recent decades, the development of new drugs has become increasingly expensive and inefficient, and the molecular mechanisms of most pharmaceuticals remain poorly understood. In response, computational systems and network medicine tools have emerged to identify potential drug repurposing candidates. However, these tools often require complex installation and lack intuitive visual network mining capabilities. To tackle these challenges, we introduce Drugst.One, a platform that assists specialized computational medicine tools in becoming user-friendly, web-based utilities for drug repurposing. With just three lines of code, Drugst.One turns any systems biology software into an interactive web tool for modeling and analyzing complex protein-drug-disease networks. Demonstrating its broad adaptability, Drugst.One has been successfully integrated with 21 computational systems medicine tools. Available at https://drugst.one, Drugst.One has significant potential for streamlining the drug discovery process, allowing researchers to focus on essential aspects of pharmaceutical treatment research.

Online bias-aware disease module mining with ROBUST-Web.

SUMMARY: We present ROBUST-Web which implements our recently presented ROBUST disease module mining algorithm in a user-friendly web application. ROBUST-Web features seamless downstream disease module exploration via integrated gene set enrichment analysis, tissue expression annotation, and visualization of drug-protein and disease-gene links. Moreover, ROBUST-Web includes bias-aware edge costs for the underlying Steiner tree model as a new algorithmic feature, which allow to correct for study bias in protein-protein interaction networks and further improves the robustness of the computed modules. AVAILABILITY AND IMPLEMENTATION: Web application: https://robust-web.net. Source code of web application and Python package with new bias-aware edge costs: https://github.com/bionetslab/robust-web, https://github.com/bionetslab/robust_bias_aware.

spongEffects: ceRNA modules offer patient-specific insights into the miRNA regulatory landscape.

MOTIVATION: Cancer is one of the leading causes of death worldwide. Despite significant improvements in prevention and treatment, mortality remains high for many cancer types. Hence, innovative methods that use molecular data to stratify patients and identify biomarkers are needed. Promising biomarkers can also be inferred from competing endogenous RNA (ceRNA) networks that capture the gene-miRNA gene regulatory landscape. Thus far, the role of these biomarkers could only be studied globally but not in a sample-specific manner. To mitigate this, we introduce spongEffects, a novel method that infers subnetworks (or modules) from ceRNA networks and calculates patient- or sample-specific scores related to their regulatory activity. RESULTS: We show how spongEffects can be used for downstream interpretation and machine learning tasks such as tumor classification and for identifying subtype-specific regulatory interactions. In a concrete example of breast cancer subtype classification, we prioritize modules impacting the biology of the different subtypes. In summary, spongEffects prioritizes ceRNA modules as biomarkers and offers insights into the miRNA regulatory landscape. Notably, these module scores can be inferred from gene expression data alone and can thus be applied to cohorts where miRNA expression information is lacking. AVAILABILITY AND IMPLEMENTATION: https://bioconductor.org/packages/devel/bioc/html/SPONGE.html.

circRNA-sponging: a pipeline for extensive analysis of circRNA expression and their role in miRNA sponging.

MOTIVATION: Circular RNAs (circRNAs) are long non-coding RNAs (lncRNAs) often associated with diseases and considered potential biomarkers for diagnosis and treatment. Among other functions, circRNAs have been shown to act as microRNA (miRNA) sponges, preventing the role of miRNAs that repress their targets. However, there is no pipeline to systematically assess the sponging potential of circRNAs. RESULTS: We developed circRNA-sponging, a nextflow pipeline that (1) identifies circRNAs via backsplicing junctions detected in RNA-seq data, (2) quantifies their expression values in relation to their linear counterparts spliced from the same gene, (3) performs differential expression analysis, (4) identifies and quantifies miRNA expression from miRNA-sequencing (miRNA-seq) data, (5) predicts miRNA binding sites on circRNAs, (6) systematically investigates potential circRNA-miRNA sponging events, (7) creates a network of competing endogenous RNAs, and (8) identifies potential circRNA biomarkers. We showed the functionality of the circRNA-sponging pipeline using RNA sequencing data from brain tissues, where we identified two distinct types of circRNAs characterized by a specific ratio of the number of the binding site to the length of the transcript. The circRNA-sponging pipeline is the first end-to-end pipeline to identify circRNAs and their sponging systematically with raw total RNA-seq and miRNA-seq files, allowing us to better indicate the functional impact of circRNAs as a routine aspect in transcriptomic research. AVAILABILITY: https://github.com/biomedbigdata/circRNA-sponging Contact: markus.daniel.hoffmann@tum.de; markus.list@tum.de Supplementary Material: Supplementary data are available at Bioinformatic Advances online.