GitHub - linnykos/veloUncertainty Contribute to linnykos/veloUncertainty development by creating an account on GitHub.

💻 Code and analyses for this study are available at

Applied to five RNA velocity methods across mouse erythroid, pancreatic, and human brain development datasets, it also introduces a signal-to-random coherence score to guide method selection toward biologically meaningful fits.

This study introduces a replicate coherence framework for evaluating RNA velocity stability, using negative binomial count splitting to generate independent data replicates from scRNA-seq count matrices.

📊 Recently published in Bioinformatics Advances: "Quantifying stability via count splitting to guide model selection in RNA velocity analyses" 

Full text at https://doi.org/10.1093/bioadv/vbag104

GitHub - lodimk2/cortado-marker Contribute to lodimk2/cortado-marker development by creating an account on GitHub.

💻 CORTADO is available at https://github.com/lodimk2/cortado-marker

An iterative refinement procedure then uses selected markers as input features for community detection, improving clustering accuracy across brain, immune, spatial, and cancer datasets.

CORTADO uses stochastic hill-climbing optimization to select scRNA-seq marker genes by jointly maximizing differential expression, minimizing cosine similarity between selected genes, and enforcing sparsity.

🧬 New in Bioinformatics Advances: "CORTADO: Hill climbing optimization for cell-type specific marker gene discovery and clustering accuracy improvement" 

Read it at https://doi.org/10.1093/bioadv/vbag106

GitHub - GalGilad/DOMUS Contribute to GalGilad/DOMUS development by creating an account on GitHub.

💻 Code for DOMUS is available at https://github.com/GalGilad/DOMUS

DOMUS is a hierarchical clustering framework that minimizes Dasgupta's cost by blending multiple structural views of a similarity matrix via surrogate-assisted optimization. Benchmarked on synthetic, classic, & scRNA-seq datasets, it consistently outperforms average linkage & beam search baselines.

🌳 Check out the latest in Bioinformatics Advances: "An optimization framework for hierarchical clustering" 

Read it here: https://doi.org/10.1093/bioadv/vbag107

GitHub - prob-ml/spice: Deep Generative Models for Spatial Transcriptomics Deep Generative Models for Spatial Transcriptomics - prob-ml/spice

💻 Code to reproduce the results of this study is available at  https://github.com/prob-ml/spice

SPICE uses a graph convolutional network to identify genes affected by cell-cell communication in spatial transcriptomics data. Cells are represented as graph nodes connected by spatial proximity, w/ response gene expression predicted from ligand & receptor inputs across varying neighborhood radii.

🧠 Out now in Bioinformatics Advances: "Graph convolutional networks for inferring cell-cell communication from spatial transcriptomics data" 

Full text at https://doi.org/10.1093/bioadv/vbag101

GitHub - beatusmodest/AMR-in-One-health: Source code supporting population-level analysis of Escherichia coli antimicrobial resistance within a One Health contex Source code supporting population-level analysis of Escherichia coli antimicrobial resistance within a One Health contex - beatusmodest/AMR-in-One-health

💻 Analysis and source code are available at

Using ResFinder, PCA, MLST, and plasmid network analysis, it identifies cross-source resistance gene sharing and population structure. Plasmid types IncFIA, IncI1, and IncFII showed the broadest cross-source connectivity, with strongest links between human and livestock isolates.

This study characterized AMR gene distribution across 174 whole-genome sequences from human, livestock, fish, and environmental E. coli isolates in Tanzania and Kenya.

🦠 New research examines "One Health analysis of antimicrobial resistance in Escherichia coli from humans, animals, and the environment" 

Read it at https://doi.org/10.1093/bioadv/vbag099

GitHub - RosariaTornisiello/Genotype_HMM: HMM model applied to scRNA-seq data to segment chromosomes into homogenous genotype bocks HMM model applied to scRNA-seq data to segment chromosomes into homogenous genotype bocks - RosariaTornisiello/Genotype_HMM

💻 Source code, documentation, and installation instructions for scGeno can be downloaded at

scGeno infers chromosome-level genotype states from scRNA-seq data using a categorical hidden Markov model. It resolves homozygous and heterozygous chromosomal segments by modeling sequential allelic expression ratios, including crossover breakpoints, in genetically mixed single-cell datasets.

🔬 Introducing "scGeno: A hidden Markov model approach to denoise chromosome-scale genotypes from single-cell data" 

Full text at  https://doi.org/10.1093/bioadv/vbag094

Authors include: @helenekretzmer.bsky.social

🔍 Analysis of over 3.2 million prokaryotic genomes in NCBI reveals MAGs represent 91% of all available archaeal genomes, underscoring the domain's persistent cultivation challenges.

This review argues prokaryotic pangenomes are statistical models shaped by dataset quality & taxonomic resolution, not fixed biological entities. It covers pangenome fluidity, MAG-associated challenges, & the under-explored archaeal pangenome, with a look toward graph-based & AI-driven approaches.

🧬 New in Bioinformatics Advances: "The pangenome: A statistical model, not a fixed biological property" 

Read it at https://doi.org/10.1093/bioadv/vbag069

GitHub - ningzhaoAnschutz/microlive: MicroLive: An Image Processing Toolkit for Quantifying Live-cell Single-Molecule Microscopy. MicroLive: An Image Processing Toolkit for Quantifying Live-cell Single-Molecule Microscopy. - ningzhaoAnschutz/microlive

🖥️ MicroLive is available at

Validation used synthetic data from rSNAPed alongside live U-2 OS cell imaging.

MicroLive is a Python-based GUI toolkit for quantifying live-cell single-molecule microscopy data. It integrates cell segmentation, single-particle detection and tracking, colocalization, and correlation analysis into one platform.

🔬 Introducing MicroLive in Bioinformatics Advances: "MicroLive: An image processing toolkit for quantifying live-cell single-molecule microscopy" 

Full paper at https://doi.org/10.1093/bioadv/vbag095 

Authors include: @brian-munsky.bsky.social

GitHub - caoying2024-hue/peptide: MHCII-peptide MHCII-peptide. Contribute to caoying2024-hue/peptide development by creating an account on GitHub.

💻 Code and data are available at 

The model learns site-specific amino acid frequencies and pairwise joint frequencies, cross-validated by a Monte Carlo simulated annealing algorithm. Designed peptides achieved near-native binding affinities and structural confidence scores above 90, validated by DeepMHCII and AlphaFold3.