SUMMARY: We present NewWave, a scalable R/Bioconductor package for the dimensionality reduction and batch effect removal of single-cell RNA sequencing data. To achieve scalability, NewWave uses mini-batch optimization and can work with out-of-memory data, enabling users to analyze datasets with millions of cells. AVAILABILITY AND IMPLEMENTATION: NewWave is implemented as an open-source R package available through the Bioconductor project at https://bioconductor.org/packages/NewWave/. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

MOTIVATION: Metabolomics is an emerging 'omics' science involving the characterization of metabolites and metabolism in biological systems. Few bioinformatic tools have been developed for the visualization, exploration and analysis of metabolomic data within the context of metabolic pathways: some of them became rapidly obsolete and are no longer supported, others are based on a single database. A systematic collection of existing annotations has the potential of considerably boosting the investigation and contextualization of metabolomic measurements. RESULTS: We have released a major update of our Bioconductor package graphite which explicitly tracks small molecules within pathway topologies and their interactions with proteins. The package gathers the information stored in eight major databases, oriented both at genes and at metabolites, across 14 different species. Depending on user preferences, all pathways can be retrieved as gene-only, gene metabolite or metabolite-only networks. AVAILABILITY AND IMPLEMENTATION: The new graphite version (1.24) is available on Bioconductor. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

BACKGROUND: Time-course gene expression experiments are useful tools for exploring biological processes. In this type of experiments, gene expression changes are monitored along time. Unfortunately, replication of time series is still costly and usually long time course do not have replicates. Many approaches have been proposed to deal with this data structure, but none of them in the field of pathway analysis. Pathway analyses have acquired great relevance for helping the interpretation of gene expression data. Several methods have been proposed to this aim: from the classical enrichment to the more complex topological analysis that gains power from the topology of the pathway. None of them were devised to identify temporal variations in time course data. RESULTS: Here we present timeClip, a topology based pathway analysis specifically tailored to long time series without replicates. timeClip combines dimension reduction techniques and graph decomposition theory to explore and identify the portion of pathways that is most time-dependent. In the first step, timeClip selects the time-dependent pathways; in the second step, the most time dependent portions of these pathways are highlighted. We used timeClip on simulated data and on a benchmark dataset regarding mouse muscle regeneration model. Our approach shows good performance on different simulated settings. On the real dataset, we identify 76 time-dependent pathways, most of which known to be involved in the regeneration process. Focusing on the 'mTOR signaling pathway' we highlight the timing of key processes of the muscle regeneration: from the early pathway activation through growth factor signals to the late burst of protein production needed for the fiber regeneration. CONCLUSIONS: timeClip represents a new improvement in the field of time-dependent pathway analysis. It allows to isolate and dissect pathways characterized by time-dependent components. Furthermore, using timeClip on a mouse muscle regeneration dataset we were able to characterize the process of muscle fiber regeneration with its correct timing.

BACKGROUND: Gene set analysis is moving towards considering pathway topology as a crucial feature. Pathway elements are complex entities such as protein complexes, gene family members and chemical compounds. The conversion of pathway topology to a gene/protein networks (where nodes are a simple element like a gene/protein) is a critical and challenging task that enables topology-based gene set analyses.Unfortunately, currently available R/Bioconductor packages provide pathway networks only from single databases. They do not propagate signals through chemical compounds and do not differentiate between complexes and gene families. RESULTS: Here we present graphite, a Bioconductor package addressing these issues. Pathway information from four different databases is interpreted following specific biologically-driven rules that allow the reconstruction of gene-gene networks taking into account protein complexes, gene families and sensibly removing chemical compounds from the final graphs. The resulting networks represent a uniform resource for pathway analyses. Indeed, graphite provides easy access to three recently proposed topological methods. The graphite package is available as part of the Bioconductor software suite. CONCLUSIONS: graphite is an innovative package able to gather and make easily available the contents of the four major pathway databases. In the field of topological analysis graphite acts as a provider of biological information by reducing the pathway complexity considering the biological meaning of the pathway elements.

MOTIVATION: Inferring large transcriptional networks using mutual information has been shown to be effective in several experimental setup. Unfortunately, this approach has two main drawbacks: (i) several mutual information estimators are prone to biases and (ii) available software still has large computational costs when processing thousand of genes. RESULTS: Here, we present parmigene (PARallel Mutual Information estimation for GEne NEtwork reconstruction), an R package that tries to fill the above gaps. It implements a mutual information estimator based on k-nearest neighbor distances that is minimally biased with respect to the other methods and uses a parallel computing paradigm to reconstruct gene regulatory networks. We test parmigene on in silico and real data. We show that parmigene gives more precise results than existing softwares with strikingly less computational costs. AVAILABILITY AND IMPLEMENTATION: The parmigene package is available on the CRAN network at http://cran.r-project.org/web/packages/. CONTACT: [email protected]

BACKGROUND: In the last decades, microarray technology has spread, leading to a dramatic increase of publicly available datasets. The first statistical tools developed were focused on the identification of significant differentially expressed genes. Later, researchers moved toward the systematic integration of gene expression profiles with additional biological information, such as chromosomal location, ontological annotations or sequence features. The analysis of gene expression linked to physical location of genes on chromosomes allows the identification of transcriptionally imbalanced regions, while, Gene Set Analysis focuses on the detection of coordinated changes in transcriptional levels among sets of biologically related genes. In this field, meta-analysis offers the possibility to compare different studies, addressing the same biological question to fully exploit public gene expression datasets. RESULTS: We describe STEPath, a method that starts from gene expression profiles and integrates the analysis of imbalanced region as an a priori step before performing gene set analysis. The application of STEPath in individual studies produced gene set scores weighted by chromosomal activation. As a final step, we propose a way to compare these scores across different studies (meta-analysis) on related biological issues. One complication with meta-analysis is batch effects, which occur because molecular measurements are affected by laboratory conditions, reagent lots and personnel differences. Major problems occur when batch effects are correlated with an outcome of interest and lead to incorrect conclusions. We evaluated the power of combining chromosome mapping and gene set enrichment analysis, performing the analysis on a dataset of leukaemia (example of individual study) and on a dataset of skeletal muscle diseases (meta-analysis approach). In leukaemia, we identified the Hox gene set, a gene set closely related to the pathology that other algorithms of gene set analysis do not identify, while the meta-analysis approach on muscular disease discriminates between related pathologies and correlates similar ones from different studies. CONCLUSIONS: STEPath is a new method that integrates gene expression profiles, genomic co-expressed regions and the information about the biological function of genes. The usage of the STEPath-computed gene set scores overcomes batch effects in the meta-analysis approaches allowing the direct comparison of different pathologies and different studies on a gene set activation level.