1012 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.
PerturbLDM is a conditional latent diffusion framework pretrained on the Tahoe-100M dataset that outperforms existing methods in accurately predicting and generating single-cell transcriptional responses to unseen drug, dose, and cell line combinations across diverse biological contexts.
This paper demonstrates that applying conformal prediction to the BayesAge epigenetic clock enables the generation of distribution-free, calibrated prediction intervals for DNA methylation-based age estimation using a minimal set of CpG sites, thereby addressing the lack of uncertainty quantification in existing point-estimate models.
This paper demonstrates that cryptic binding site detection is currently limited not by the ability to propose candidates, but by poor ranking and lack of consensus, revealing that separating coverage from conversion metrics uncovers significant performance gaps and shows that combining geometric and language-model-based detectors within a strict candidate budget yields the most practical improvements.
The authors introduce ProtInt, a deep learning framework that successfully integrates proteomic data from cancer cell lines and patient tumors by overcoming missing value challenges, thereby outperforming existing methods and revealing key biological shifts needed to better align preclinical models with clinical reality.
This systematic benchmarking study reveals that the choice of computational method, rather than underlying biology, is the primary determinant of in silico perturbation results, as most widely used approaches fail to detect causal transcription factor-to-pathway signals and often produce conclusions that contradict experimental CRISPRi validation.
This paper demonstrates that cross-modal pretraining of a single-cell foundation model with proteomics data yields superior gene-level and cell-level representations and better generalization than simply scaling up RNA-only models, highlighting the value of multimodal training over parameter scaling alone.
This paper introduces "sigRecon," a comprehensive benchmarking framework and open-source R package that systematically evaluates methods for predicting perturbation signatures across diverse biological contexts, revealing that projection and network-based approaches often match or outperform complex deep learning models while highlighting key factors influencing cross-context generalization.
Cell-Hub is a free, open-source, Docker-based graphical interface built on R/Shiny that democratizes end-to-end single-cell RNA sequencing analysis by integrating advanced tools like Seurat 5, CellChat 2, and Monocle 3 into a unified, no-code platform for researchers of all computational backgrounds.
This study demonstrates that using IUPAC consensus references with the ambiguity-aware aligner novoAlign significantly improves short-read variant detection in clinically challenging genomic regions compared to standard linear references, while introducing the open-source BurdenBench framework to better evaluate caller-specific precision-recall trade-offs and net clinical benefit.
The paper introduces "disize," a novel RNA-seq normalization method that leverages experimental design information through a modified generalized linear mixed model to more accurately estimate size factors and improve downstream differential expression analysis, particularly in challenging scenarios with widespread expression changes.