1180 papers
Bioinformatics sits at the exciting intersection where biology meets data science, using powerful computer tools to decode the vast complexity of living systems. From mapping the human genome to tracking how viruses evolve, this field transforms raw biological information into actionable insights that drive modern medicine and research forward without requiring a supercomputer to understand the basics.
On Gist.Science, we ensure you never miss a breakthrough by processing every new preprint in this category directly from bioRxiv. Our team provides both plain-language explanations and detailed technical summaries for each paper, making cutting-edge discoveries accessible to everyone regardless of their background.
Below are the latest bioinformatics papers added from bioRxiv, ready for you to explore with clarity and depth.
VirtualSkin is a cell language model that successfully integrates skin cell identity with intervention-response prediction, demonstrating improved accuracy in annotating inflammatory skin states and forecasting cellular responses across diverse diseases, anatomical sites, and perturbations.
This study introduces the LEAF framework to evaluate whether language models improve budgeted molecular search for ionizable lipids, finding that despite varying degrees of evidence engagement, zero-shot LLM proposers fail to outperform classical baselines like genetic algorithms due to high duplication rates and a lack of genuine performance gains.
The study introduces EcoNet, a deep-learning framework that models structured tumor microenvironment ecosystems and signaling networks to significantly outperform existing methods in predicting immunotherapy response, as validated by its ability to identify biologically relevant gene programs and stratify clinical outcomes in clear cell renal cell carcinoma.
BioMod is a 50-million-parameter single-cell foundation model that improves representation efficiency and performance by tokenizing cells into biologically meaningful overlapping modules rather than individual genes, achieving state-of-the-art results in tasks like cell-type annotation and perturbation prediction while enabling efficient inference on a single GPU.
This paper introduces an explainable Gaussian random field framework for multiple imputation of longitudinal compositional data that effectively handles missing samples, temporal dependencies, and zero inflation, demonstrating its utility in reconstructing gut microbiome trajectories to improve predictions for food allergy and postpartum depression.
AmyloCore-ML is a novel sequence-based predictor that leverages protein language model embeddings, particularly ESM-2, to accurately identify structurally resolved amyloid fibril core regions, outperforming existing tools like CrossBeta and AggrescanAI.
This study presents MtbAIM, an AI/ML-driven platform that utilizes diverse molecular representations and tree-based algorithms to accurately predict the activity and minimum inhibitory concentrations of small molecules against both drug-susceptible and drug-resistant *Mycobacterium tuberculosis* isolates.
The authors present an enhanced, alignment-free version of the Xengsort tool that enables efficient and accurate separation of DNA reads by species for epigenetic profiling of patient-derived xenograft models, outperforming existing methods in computational efficiency.
This paper revisits a 1998 model for human intestinal absorption by demonstrating that its original performance relied on a fortunate data split, and subsequently establishes more robust, consensus-based predictive models for intestinal absorption and various permeability assays using genetic algorithm-selected descriptor sets and random forests across 250 stratified extractions.
GenDiMug is a novel multigraph framework that overcomes the limitations of linear genome representations by using network models to integrate multi-omics data and analyze genomic organization across multiple scales, revealing structural patterns such as gene expression clumping and circadian regulation.