BACKGROUND: Rapid extraction and visualization of cell-specific gene expression is important for automatic cell type annotation, e.g. in single cell analysis. There is an emerging field in which tools such as curated databases or machine learning methods are used to support cell type annotation. However, complementing approaches to efficiently incorporate the latest knowledge of free-text articles from literature databases, such as PubMed, are understudied. RESULTS: This work introduces the PubMed Gene/Cell type-Relation Atlas (PuMA) which provides a local, easy-to-use web-interface to facilitate literature-driven cell type annotation. It utilizes a pretrained machine learning based named entity recognition model in order to extract gene and cell type concepts from PubMed, links biomedical ontologies, and suggests gene to cell type relations based on a ranking score. It includes a search tool for genes and cell types, additionally providing an interactive graph visualization for exploring cross-relations. Each result is fully traceable by linking the relevant PubMed articles. CONCLUSIONS: This work enables researchers to analyse and automatize cell type annotation based on PubMed articles. It complements manual curated marker gene databases and enables interactive visualizations. The evaluation shows that PuMA is competitive against an extensive manual curated database across three gold standard datasets and two species-mouse and human. The software framework is freely available and enables regular article imports for incremental knowledge updates.GitLab: https://imigitlab.uni-muenster.de/published/PuMA/.

MOTIVATION: With its capacity for high-resolution data output in one region of interest, chromosome conformation capture combined with high-throughput sequencing (4C-seq) is a state-of-the-art next-generation sequencing technique that provides epigenetic insights, and regularly advances current medical research. However, 4C-seq data are complex and prone to biases, and while specialized programs exist, an unbiased, extensive benchmarking is still lacking. Furthermore, neither substantial datasets with fully characterized ground truth, nor simulation programs for realistic 4C-seq data have been published. RESULTS: We conducted a benchmarking study on 66 4C-seq samples from 20 datasets, and developed a novel 4C-seq simulation software, Basic4CSim, to allow for detailed comparisons of 4C-seq algorithms on 50 simulated datasets with 10-120 samples each. Simulations and benchmarking were adapted to address different characteristics of 4C-seq data. Simulated data were compared with published samples to validate simulation settings. We identified differences between 4C-seq algorithms in terms of precision, recall, interaction structure, and run time, and observed general trends. Novel differential pipeline versions of single-sample based 4C-seq algorithms were included in the benchmarking. While no single tool was optimally suited for both near-cis and far-cis, and both single-sample and differential analyses, choosing a high-performing algorithm variant did improve results considerably. For near-cis scenarios, r3Cseq, peakC and FourCSeq offered high precision, while fourSig demonstrated high overall F1 scores in far-cis analyses. Finally, 4C-seq simulations may aid in the development of improved analysis algorithms. AVAILABILITY AND IMPLEMENTATION: Basic4CSim is available at https://github.com/walter-ca/Basic4CSim. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

SUMMARY: Basic4Cseq is an R/Bioconductor package for basic filtering, analysis and subsequent near-cis visualization of 4C-seq data. The package processes aligned 4C-seq raw data stored in binary alignment/map (BAM) format and maps the short reads to a corresponding virtual fragment library. Functions are included to create virtual fragment libraries providing chromosome position and further information on 4C-seq fragments (length and uniqueness of the fragment ends, and blindness of a fragment) for any BSGenome package. An optional filter is included for BAM files to remove invalid 4C-seq reads, and further filter functions are offered for 4C-seq fragments. Additionally, basic quality controls based on the read distribution are included. Fragment data in the vicinity of the experiment's viewpoint are visualized as coverage plot based on a running median approach and a multi-scale contact profile. Wig files or csv files of the fragment data can be exported for further analyses and visualizations of interactions with other programs. AVAILABILITY AND IMPLEMENTATION: Basic4Cseq is implemented in R and available at http://www.bioconductor.org/. A vignette with detailed descriptions of the functions is included in the package. CONTACT: [email protected] SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.