MOTIVATION: Sequence-based deep learning approaches have been shown to predict a multitude of functional genomic readouts, including regions of open chromatin and RNA expression of genes. However, a major limitation of current methods is that model interpretation relies on computationally demanding post hoc analyses, and even then, one can often not explain the internal mechanics of highly parameterized models. Here, we introduce a deep learning architecture called totally interpretable sequence-to-function model (tiSFM). tiSFM improves upon the performance of standard multilayer convolutional models while using fewer parameters. Additionally, while tiSFM is itself technically a multilayer neural network, internal model parameters are intrinsically interpretable in terms of relevant sequence motifs. RESULTS: We analyze published open chromatin measurements across hematopoietic lineage cell-types and demonstrate that tiSFM outperforms a state-of-the-art convolutional neural network model custom-tailored to this dataset. We also show that it correctly identifies context-specific activities of transcription factors with known roles in hematopoietic differentiation, including Pax5 and Ebf1 for B-cells, and Rorc for innate lymphoid cells. tiSFM's model parameters have biologically meaningful interpretations, and we show the utility of our approach on a complex task of predicting the change in epigenetic state as a function of developmental transition. AVAILABILITY AND IMPLEMENTATION: The source code, including scripts for the analysis of key findings, can be found at https://github.com/boooooogey/ATAConv, implemented in Python.

SUMMARY: DNA changes that cause premature termination codons (PTCs) represent a large fraction of clinically relevant pathogenic genomic variation. Typically, PTCs induce transcript degradation by nonsense-mediated mRNA decay (NMD) and render such changes loss-of-function alleles. However, certain PTC-containing transcripts escape NMD and can exert dominant-negative or gain-of-function (DN/GOF) effects. Therefore, systematic identification of human PTC-causing variants and their susceptibility to NMD contributes to the investigation of the role of DN/GOF alleles in human disease. Here we present aenmd, a software for annotating PTC-containing transcript-variant pairs for predicted escape from NMD. aenmd is user-friendly and self-contained. It offers functionality not currently available in other methods and is based on established and experimentally validated rules for NMD escape; the software is designed to work at scale, and to integrate seamlessly with existing analysis workflows. We applied aenmd to variants in the gnomAD, Clinvar, and GWAS catalog databases and report the prevalence of human PTC-causing variants in these databases, and the subset of these variants that could exert DN/GOF effects via NMD escape. AVAILABILITY AND IMPLEMENTATION: aenmd is implemented in the R programming language. Code is available on GitHub as an R-package (github.com/kostkalab/aenmd.git), and as a containerized command-line interface (github.com/kostkalab/aenmd_cli.git).

MOTIVATION: Single-cell RNA sequencing (scRNA-seq) continues to expand our knowledge by facilitating the study of transcriptional heterogeneity at the level of single cells. Despite this technology's utility and success in biomedical research, technical artifacts are present in scRNA-seq data. Doublets/multiplets are a type of artifact that occurs when two or more cells are tagged by the same barcode, and therefore they appear as a single cell. Because this introduces non-existent transcriptional profiles, doublets can bias and mislead downstream analysis. To address this limitation, computational methods to annotate and remove doublets form scRNA-seq datasets are needed. RESULTS: We introduce vaeda (Variational Auto-Encoder for Doublet Annotation), a new approach for computational annotation of doublets in scRNA-seq data. Vaeda integrates a variational auto-encoder and Positive-Unlabeled learning to produce doublet scores and binary doublet calls. We apply vaeda, along with seven existing doublet annotation methods, to 16 benchmark datasets and find that vaeda performs competitively in terms of doublet scores and doublet calls. Notably, vaeda outperforms other python-based methods for doublet annotation. Altogether, vaeda is a robust and competitive method for scRNA-seq doublet annotation and may be of particular interest in the context of python-based workflows. AVAILABILITY AND IMPLEMENTATION: Vaeda is available at https://github.com/kostkalab/vaeda, and the version used for the results we present here is archived at zenodo (https://doi.org/10.5281/zenodo.7199783). SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

MOTIVATION: Single-cell RNA-seq analysis has emerged as a powerful tool for understanding inter-cellular heterogeneity. Due to the inherent noise of the data, computational techniques often rely on dimensionality reduction (DR) as both a pre-processing step and an analysis tool. Ideally, DR should preserve the biological information while discarding the noise. However, if the DR is to be used directly to gain biological insight it must also be interpretable-that is the individual dimensions of the reduction should correspond to specific biological variables such as cell-type identity or pathway activity. Maximizing biological interpretability necessitates making assumption about the data structures and the choice of the model is critical. RESULTS: We present a new probabilistic single-cell factor analysis model, Non-negative Independent Factor Analysis (NIFA), that incorporates different interpretability inducing assumptions into a single modeling framework. The key advantage of our NIFA model is that it simultaneously models uni- and multi-modal latent factors, and thus isolates discrete cell-type identity and continuous pathway activity into separate components. We apply our approach to a range of datasets where cell-type identity is known, and we show that NIFA-derived factors outperform results from ICA, PCA, NMF and scCoGAPS (an NMF method designed for single-cell data) in terms of disentangling biological sources of variation. Studying an immunotherapy dataset in detail, we show that NIFA is able to reproduce and refine previous findings in a single analysis framework and enables the discovery of new clinically relevant cell states. AVAILABILITY AND IMPLEMENTATION: NFIA is a R package which is freely available at GitHub (https://github.com/wgmao/NIFA). The test dataset is archived at https://zenodo.org/record/6286646. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

MOTIVATION: Single-cell RNA sequencing (scRNA-seq) technologies enable the study of transcriptional heterogeneity at the resolution of individual cells and have an increasing impact on biomedical research. However, it is known that these methods sometimes wrongly consider two or more cells as single cells, and that a number of so-called doublets is present in the output of such experiments. Treating doublets as single cells in downstream analyses can severely bias a study's conclusions, and therefore computational strategies for the identification of doublets are needed. RESULTS: With scds, we propose two new approaches for in silico doublet identification: Co-expression based doublet scoring (cxds) and binary classification based doublet scoring (bcds). The co-expression based approach, cxds, utilizes binarized (absence/presence) gene expression data and, employing a binomial model for the co-expression of pairs of genes, yields interpretable doublet annotations. bcds, on the other hand, uses a binary classification approach to discriminate artificial doublets from original data. We apply our methods and existing computational doublet identification approaches to four datasets with experimental doublet annotations and find that our methods perform at least as well as the state of the art, at comparably little computational cost. We observe appreciable differences between methods and across datasets and that no approach dominates all others. In summary, scds presents a scalable, competitive approach that allows for doublet annotation of datasets with thousands of cells in a matter of seconds. AVAILABILITY AND IMPLEMENTATION: scds is implemented as a Bioconductor R package (doi: 10.18129/B9.bioc.scds). SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

MOTIVATION: MicroRNAs (miRNAs) are important non-coding post-transcriptional regulators that are involved in many biological processes and human diseases. Individual miRNAs may regulate hundreds of genes, giving rise to a complex gene regulatory network in which transcripts carrying miRNA binding sites act as competing endogenous RNAs (ceRNAs). Several methods for the analysis of ceRNA interactions exist, but these do often not adjust for statistical confounders or address the problem that more than one miRNA interacts with a target transcript. RESULTS: We present SPONGE, a method for the fast construction of ceRNA networks. SPONGE uses 'multiple sensitivity correlation', a newly defined measure for which we can estimate a distribution under a null hypothesis. SPONGE can accurately quantify the contribution of multiple miRNAs to a ceRNA interaction with a probabilistic model that addresses previously neglected confounding factors and allows fast P-value calculation, thus outperforming existing approaches. We applied SPONGE to paired miRNA and gene expression data from The Cancer Genome Atlas for studying global effects of miRNA-mediated cross-talk. Our results highlight already established and novel protein-coding and non-coding ceRNAs which could serve as biomarkers in cancer. AVAILABILITY AND IMPLEMENTATION: SPONGE is available as an R/Bioconductor package (doi: 10.18129/B9.bioc.SPONGE). SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Motivation: Genome-wide transcriptome sequencing applied to single cells (scRNA-seq) is rapidly becoming an assay of choice across many fields of biological and biomedical research. Scientific objectives often revolve around discovery or characterization of types or sub-types of cells, and therefore, obtaining accurate cell-cell similarities from scRNA-seq data is a critical step in many studies. While rapid advances are being made in the development of tools for scRNA-seq data analysis, few approaches exist that explicitly address this task. Furthermore, abundance and type of noise present in scRNA-seq datasets suggest that application of generic methods, or of methods developed for bulk RNA-seq data, is likely suboptimal. Results: Here, we present RAFSIL, a random forest based approach to learn cell-cell similarities from scRNA-seq data. RAFSIL implements a two-step procedure, where feature construction geared towards scRNA-seq data is followed by similarity learning. It is designed to be adaptable and expandable, and RAFSIL similarities can be used for typical exploratory data analysis tasks like dimension reduction, visualization and clustering. We show that our approach compares favorably with current methods across a diverse collection of datasets, and that it can be used to detect and highlight unwanted technical variation in scRNA-seq datasets in situations where other methods fail. Overall, RAFSIL implements a flexible approach yielding a useful tool that improves the analysis of scRNA-seq data. Availability and implementation: The RAFSIL R package is available at www.kostkalab.net/software.html. Supplementary information: Supplementary data are available at Bioinformatics online.

MOTIVATION: Methylation of CpG dinucleotides is a prevalent epigenetic modification that is required for proper development in vertebrates. Genome-wide DNA methylation assays have become increasingly common, and this has enabled characterization of DNA methylation in distinct stages across differentiating cellular lineages. Changes in CpG methylation are essential to cellular differentiation; however, current methods for modeling methylation dynamics do not account for the dependency structure between precursor and dependent cell types. RESULTS: We developed a continuous-time Markov chain approach, based on the observation that changes in methylation state over tissue differentiation can be modeled similarly to DNA nucleotide changes over evolutionary time. This model explicitly takes precursor to descendant relationships into account and enables inference of CpG methylation dynamics. To illustrate our method, we analyzed a high-resolution methylation map of the differentiation of mouse stem cells into several blood cell types. Our model can successfully infer unobserved CpG methylation states from observations at the same sites in related cell types (90% correct), and this approach more accurately reconstructs missing data than imputation based on neighboring CpGs (84% correct). Additionally, the single CpG resolution of our methylation dynamics estimates enabled us to show that DNA sequence context of CpG sites is informative about methylation dynamics across tissue differentiation. Finally, we identified genomic regions with clusters of highly dynamic CpGs and present a likely functional example. Our work establishes a framework for inference and modeling that is well suited to DNA methylation data, and our success suggests that other methods for analyzing DNA nucleotide substitutions will also translate to the modeling of epigenetic phenomena. AVAILABILITY AND IMPLEMENTATION: Source code is available at www.kostkalab.net/software.

Gene-regulatory enhancers have been identified using various approaches, including evolutionary conservation, regulatory protein binding, chromatin modifications, and DNA sequence motifs. To integrate these different approaches, we developed EnhancerFinder, a two-step method for distinguishing developmental enhancers from the genomic background and then predicting their tissue specificity. EnhancerFinder uses a multiple kernel learning approach to integrate DNA sequence motifs, evolutionary patterns, and diverse functional genomics datasets from a variety of cell types. In contrast with prediction approaches that define enhancers based on histone marks or p300 sites from a single cell line, we trained EnhancerFinder on hundreds of experimentally verified human developmental enhancers from the VISTA Enhancer Browser. We comprehensively evaluated EnhancerFinder using cross validation and found that our integrative method improves the identification of enhancers over approaches that consider a single type of data, such as sequence motifs, evolutionary conservation, or the binding of enhancer-associated proteins. We find that VISTA enhancers active in embryonic heart are easier to identify than enhancers active in several other embryonic tissues, likely due to their uniquely high GC content. We applied EnhancerFinder to the entire human genome and predicted 84,301 developmental enhancers and their tissue specificity. These predictions provide specific functional annotations for large amounts of human non-coding DNA, and are significantly enriched near genes with annotated roles in their predicted tissues and lead SNPs from genome-wide association studies. We demonstrate the utility of EnhancerFinder predictions through in vivo validation of novel embryonic gene regulatory enhancers from three developmental transcription factor loci. Our genome-wide developmental enhancer predictions are freely available as a UCSC Genome Browser track, which we hope will enable researchers to further investigate questions in developmental biology.

UNLABELLED: Nested effects models (NEMs) are a class of probabilistic models introduced to analyze the effects of gene perturbation screens visible in high-dimensional phenotypes like microarrays or cell morphology. NEMs reverse engineer upstream/downstream relations of cellular signaling cascades. NEMs take as input a set of candidate pathway genes and phenotypic profiles of perturbing these genes. NEMs return a pathway structure explaining the observed perturbation effects. Here, we describe the package nem, an open-source software to efficiently infer NEMs from data. Our software implements several search algorithms for model fitting and is applicable to a wide range of different data types and representations. The methods we present summarize the current state-of-the-art in NEMs. AVAILABILITY: Our software is written in the R language and freely avail-able via the Bioconductor project at http://www.bioconductor.org.

MOTIVATION: Molecular diagnostics aims at classifying diseases into clinically relevant sub-entities based on molecular characteristics. Typically, the entities are split into subgroups, which might contain several variants yielding a hierarchical model of the disease. Recent years have introduced a plethora of new molecular screening technologies to molecular diagnostics. As a result molecular profiles of patients became complex and the classification task more difficult. RESULTS: We present a novel tool for detecting hierarchical structure in binary datasets. We aim for identifying molecular characteristics, which are stochastically implying other characteristics. The final hierarchical structure is encoded in a directed transitive graph where nodes represent molecular characteristics and a directed edge from a node A to a node B denotes that almost all cases with characteristic B also display characteristic A. Naturally, these graphs need to be transitive. In the core of our modeling approach lies the problem of calculating good transitive approximations of given directed but not necessarily transitive graphs. By good transitive approximation we understand transitive graphs, which differ from the reference graph in only a small number of edges. It is known that the problem of finding optimal transitive approximation is NP-complete. Here we develop an efficient heuristic for generating good transitive approximations. We evaluate the computational efficiency of the algorithm in simulations, and demonstrate its use in the context of a large genome-wide study on mature aggressive lymphomas. AVAILABILITY: The software used in our analysis is freely available from http://compdiag.uni-regensburg.de/software/transApproxs.shtml.

Microarray gene expression signatures hold great promise to improve diagnosis and prognosis of disease. However, current documentation standards of such signatures do not allow for an unambiguous application to study-external patients. This hinders independent evaluation, effectively delaying the use of signatures in clinical practice. Data from eight publicly available clinical microarray studies were analyzed and the consistency of study-internal with study-external diagnoses was evaluated. Study-external classifications were based on documented information only. Documenting a signature is conceptually different from reporting a list of genes. We show that even the exact quantitative specification of a classification rule alone does not define a signature unambiguously. We found that discrepancy between study-internal and study-external diagnoses can be as frequent as 30% (worst case) and 18% (median). By using the proposed documentation by value strategy, which documents quantitative preprocessing information, the median discrepancy was reduced to 1%. The process of evaluating microarray gene expression diagnostic signatures and bringing them to clinical practice can be substantially improved and made more reliable by better documentation of the signatures.