Bioinformatics Advances

⏱️ New research in Bioinformatics Advances: “Cis-TERN: A graph-based method to integrate single-cell chromatin accessibility with gene expression data to infer the pseudotime ordering of cells” Access it here: https://doi.org/10.1093/bioadv/vbag313 Authors include: @pfprzytycki.bsky.social

Across healthy hematopoiesis and mixed-phenotype leukemia samples, its scores correlated strongly with Monocle 3 and could trace cancer cells back to their likely progenitors.

github.comGitHub - pawelab/Cis-TERN: Cis-TERN: a graph-based method to integrate single-cell chromatin accessibility with gene expression data to infer the pseudotime ordering of cellsCis-TERN: a graph-based method to integrate single-cell chromatin accessibility with gene expression data to infer the pseudotime ordering of cells - pawelab/Cis-TERN

💻 Cis-TERN is available at

Cis-TERN is a graph-based method that builds a directed bipartite graph linking scRNA-seq and scATAC-seq cells, using precedence scores derived from distal cis-regulatory accessibility to infer continuous pseudotime via network diffusion, without requiring a specified root cell type.

🔄 New in Bioinformatics Advances: “pycircdb: Integrated circRNA database annotation for computational workflows” Find it here: https://doi.org/10.1093/bioadv/vbag311

Built on Apache Hamilton and Parquet for parallelized, low-disk queries, it annotated over 97% of high-confidence candidates and processed 300,000 circRNAs in 66 seconds at a peak of 3.8 GB RAM.

github.comGitHub - BarryDigby/pycircdb: pycircdb: integrated circRNA database annotation for computational workflowspycircdb: integrated circRNA database annotation for computational workflows - BarryDigby/pycircdb

💻 pycircdb is available via PyPI at https://pypi.org/project/pycircdb under the GPL-3.0 license. Source code at

🔄 New in Bioinformatics Advances: “pycircdb: Integrated circRNA database annotation for computational workflows” Find it here: https://doi.org/10.1093/bioadv/vbag311

pycircdb is a Python command-line tool that normalizes nine curated circRNA databases to a common 0-based coordinate system, cross-mapped between hg19 and hg38, resolving conflicting identifiers through tolerant coordinate matching.

🦠 New study in Bioinformatics Advances: “Transcriptional profiling and co-expression integration for the filtering of relevant bacterial sRNA–mRNA interactions: Application to Staphylococcus aureus biofilm” Read it here: https://doi.org/10.1093/bioadv/vbag309

The pipeline narrowed 1.73 million candidate interactions to 1,290 prioritized pairs, including ten sRNAs consistently upregulated and four consistently downregulated in biofilm across all strains.

github.comGitHub - carolina5massena/Integrative-RNA-Seq-and-Co-expression-Analyses-Uncover-Confident-sRNAContribute to carolina5massena/Integrative-RNA-Seq-and-Co-expression-Analyses-Uncover-Confident-sRNA development by creating an account on GitHub.

🖥️ The open-source graphical application and analysis code are available at

This integrative framework combines sequence-based sRNA–mRNA prediction from four tools with differential-expression and co-expression network evidence to prioritize regulatory candidates across six S. aureus strains.

⚡ New paper in Bioinformatics Advances: “TorchGWAS 1.0: GPU-accelerated GWAS at scale” See it here: https://doi.org/10.1093/bioadv/vbag310

On an NVIDIA H100, it completed 4.57 billion associations across 8.9 million variants, 512 phenotypes, and 35,365 samples in 28–59 seconds, matching PLINK 2.0’s test statistics to within 1.3×10⁻¹².

github.comGitHub - ZhiGroup/TorchGWAS: GPU-accelerated GWAS for ultra-high dimensional phenotypesGPU-accelerated GWAS for ultra-high dimensional phenotypes - ZhiGroup/TorchGWAS

💻 TorchGWAS is available at

⚡ New paper in Bioinformatics Advances: “TorchGWAS 1.0: GPU-accelerated GWAS at scale” See it here: https://doi.org/10.1093/bioadv/vbag310

TorchGWAS is a GPU-accelerated framework that uses batched operations and an overlapped streaming pipeline for covariate-adjusted linear GWAS across large phenotype panels.

🧮 Recently published in Bioinformatics Advances: “1100 synthetic benchmark problems for dynamic modeling of cellular processes ” Read more: https://doi.org/10.1093/bioadv/vbag307

The synthetic problems reproduce the statistical characteristics of their templates while spanning a broader, more continuous range of optimization difficulty and parameter identifiability.

doi.org1100 Synthetic Benchmark Problems for Dynamic Modeling of Cellular Processes (Repository Archive)This archive contains the publication version of the repository 1100SyntheticBenchmarksODE associated with the manuscript: Neubrand et al. 1100 Synthetic Benchmark Problems for Dynamic Modeling of Cellular Processes. Bioinformatics Advances. The repository contains: the algorithm for generating new...

💾 Benchmark problems and the generation algorithm are available at https://github.com/niklasneubrand/1100SyntheticBenchmarksODE, archived at

🧮 Recently published in Bioinformatics Advances: “1100 synthetic benchmark problems for dynamic modeling of cellular processes ” Read more: https://doi.org/10.1093/bioadv/vbag307

This collection extends 22 published ODE models into 1100 synthetic benchmark problems by perturbing dynamic parameters and generating realistic observable structures and measurement patterns across experimental conditions.

🧬 New in Bioinformatics Advances: “RT-SRFM: A transport-growth-fate latent dynamical framework for developmental single-cell time series ” Access it here: https://doi.org/10.1093/bioadv/vbag293

On zebrafish and sea urchin datasets, it outperformed optimal-transport and neural-dynamics baselines in mass calibration and fate prediction.

github.comGitHub - typical666/RT-SRFM at clean-uploadContribute to typical666/RT-SRFM development by creating an account on GitHub.

💻 RT-SRFM’s source code and datasets are available at

RT-SRFM is a latent dynamical framework that jointly models single-cell transport, growth, and fate in one shared space, trained via an eight-stage curriculum combining OT-guided velocity learning and multi-step rollout repair.

In Memoriam in Bioinformatics Advances: “Peer Bork (1963–2026): Pioneer of computational biology, architect of bioinformatics resources, and visionary microbiome researcher” Read the tribute here: https://doi.org/10.1093/bioadv/vbag227 Author: @alexbateman1.bsky.social

doi.orgdoi.org

Over three decades, Peer pioneered protein domain discovery (WW, HEAT, BRCT), built foundational resources including SMART, STRING, eggNOG, and iTOL, and transformed microbiome science with the 2011 discovery of human gut enterotypes, cited nearly 10,000 times.

peerbork.muchloved.comIn remembrance of Peer Bork4 May 1963 - 16 January 2026

His influence spans more than 600 publications, a citation count exceeding 460,000, and over 50 trainees who now lead their own research groups. Tributes can be shared at https://peerbork.muchloved.com/.

🔬 New study in Bioinformatics Advances: "Bridging LLM-based agents and workflow languages for automated multi-perspective analysis of scRNA-seq data" Find it here: https://doi.org/10.1093/bioadv/vbag301

Validation raised the executable analysis intent fulfillment rate from 0.061 to 0.378, with parallel execution achieving up to a 29.9-fold speedup over sequential runs.

github.comGitHub - bioinfo-tsukuba/mpdas_paperContribute to bioinfo-tsukuba/mpdas_paper development by creating an account on GitHub.

💻 Source code for the framework is available at https://github.com/bioinfo-tsukuba/mpdas_paper under the MIT license.

This framework pairs LLM agents with Snakemake workflows, generating diverse scRNA-seq analytical strategies and validating each as a plan before parallel execution.

Bioinformatics Advances reposted

Alexis Stamatakis

@stamatak.bsky.social

Attend "Future of Sustainable Computing Symposium" free registration & remote participation: thefutureofsuscomp2026.h-its.org Speakers: Ivona Brandic Florina M. Ciorba Denisa Constantinescu Sophie Cerf Thaleia Dimitra Doudali Loïc Lannelongue Matteo Mendula August Ning Emma Strubell Klervie Toczé

🧬 New paper in Bioinformatics Advances: "DistPCA: Tera-scale genomic PCA via out-of-core distributed parallelism" Read it here: https://doi.org/10.1093/bioadv/vbag303

On a 500K-genome dataset it cut wall-clock time from 12.1 to 2.3 hours against the state-of-the-art PCAone, a 5.26x speedup, while preserving near-identical PC accuracy.

doi.orgDistPCA: Tera-Scale Genomic PCA via Out-of-Core Distributed ParallelismDistPCA is a distributed out-of-core C++ framework for tera-scale genomic Principal Component Analysis (PCA), designed to scale efficiently across both single- and multi-node computing systems. Built on top of Message Passing Interface (MPI), it employs a hybrid multi-level parallelism scheme combining multiprocessing, OpenMP multithreading, SIMD vectorization, and double buffering across all three stages of the PCA pipeline (I/O, data preprocessing, numerical method). Evaluated on datasets reaching up to 11 TB, DistPCA achieves speedups of up to 58.2× and over 98% reduction in wall-clock time, while maintaining parallel efficiency above 82% and preserving the accuracy of the recovered principal components (PCs). For a detailed description of the framework and experimental evaluation, please refer to our preprint. Detailed instructions on how to set up and use the framework, as well as how to reproduce our experiments, are provided in the README.md file.

🖥️ DistPCA is available at https://github.com/CEID-HPCLAB/DistPCA, with an archived release at

🧬 New paper in Bioinformatics Advances: "DistPCA: Tera-scale genomic PCA via out-of-core distributed parallelism" Read it here: https://doi.org/10.1093/bioadv/vbag303

DistPCA is a distributed out-of-core C++ framework for tera-scale genomic PCA, combining MPI, OpenMP, and SIMD vectorization to parallelize the I/O, preprocessing, and numerical stages across multi-node clusters rather than optimizing computation alone.

🦠 Latest research in Bioinformatics Advances: "Computational characterization of three conserved pathogenicity-associated genomic regions in Fusobacterium nucleatum ATCC 25586" Explore the paper: https://doi.org/10.1093/bioadv/vbag302

Iron proved essential for growth, with complete deprivation abolishing it entirely, supporting a proposed hemolysin–iron–ROS–Hippo pathway potentially relevant to colorectal cancer progression.

doi.orgexporterdream/fn-pai-characterization: Code and supplementary materials for Computational characterization of three conserved pathogenicity-associated genomic regions in Fusobacterium nucleatum ATCC 25586 Paper.Code and supplementary materials for Computational characterization of three conserved pathogenicity-associated genomic regions in Fusobacterium nucleatum ATCC 25586 Paper. Code include integrative bioinformatics and machine learning pipeline that combined genomic island prediction (IslandViewer 4),...

💻 Code, supplementary data, and analysis scripts are available at https://github.com/exporterdream/fn-pai-characterization and Zenodo at

An integrative computational pipeline combining genomic island prediction, structural modeling, and flux balance analysis identifies three conserved regions in F. nucleatum ATCC 25586 linked to hemolysis, metabolic adaptation, and immune evasion.

🦟 Recently published in Bioinformatics Advances: "Decoding channel pore dynamics mechanism in Anopheles gambiae Orco homotetramer" See it here: https://doi.org/10.1093/bioadv/vbag295

Across varying agonist:antagonist ratios, even a single IPC molecule among four bound ligands was enough to re-close an open pore, and pore widening occurred asymmetrically along the S7b helix gate rather than through uniform radial expansion

(cont'd) insights the authors suggest could guide structure-based design of mosquito repellents targeting Orco.

🦟 Recently published in Bioinformatics Advances: "Decoding channel pore dynamics mechanism in Anopheles gambiae Orco homotetramer" See it here: https://doi.org/10.1093/bioadv/vbag295

Using molecular dynamics on homotetrameric AgamOrco models, researchers compared template-based (TBM) and AlphaFold3 (AF3) structures for agonist- and antagonist-driven channel gating.

Only the TBM model reproduced experimentally observed VUAA1-induced pore opening, while the antagonist IPC consistently closed the channel and could displace agonist from an already-open pore.

🦠 New study in Bioinformatics Advances: "TSLC: Differential abundance analysis method for identifying biologically relevant known microbes" Access it here: https://doi.org/10.1093/bioadv/vbag297

It was the only method controlling false discovery at 0.05 while outperforming FishTaco in simulations.

github.comGitHub - rgbeblavy/TSLC: An R package for finding BRp taxa given functional differences.An R package for finding BRp taxa given functional differences. - rgbeblavy/TSLC

💾 TSLC is available at

🦠 New study in Bioinformatics Advances: "TSLC: Differential abundance analysis method for identifying biologically relevant known microbes" Access it here: https://doi.org/10.1093/bioadv/vbag297