Ismail, A. A. | Search bioRxiv

bioRxiv2026

CardioChrom Reconstructs the Human Cardiac Virtual Epigenome from Single-Nucleus Transcriptomes

Background: Matched epigenomic profiling remains scarce in human heart failure, limiting regulatory interpretation of single-nucleus ribonucleic acid (RNA) sequencing data. We developed CardioChrom, a cardiac-specialized framework that reconstructs chromatin accessibility, histone H3 lysine 27 acetylation (H3K27ac), and histone H3 lysine 27 trimethylation (H3K27me3) from single-nucleus RNA profiles. Methods: CardioChrom was developed using same-nucleus human heart datasets jointly measuring RNA with chromatin accessibility by assay for transposase-accessible chromatin (ATAC), H3K27ac, or H3K27me3, with donor-separated development and leave-one-cell-type-out testing. Frozen models were evaluated in 10-donor and 23-sample external human heart cohorts. Histone reconstruction was benchmarked against measured profiles and a Transformer. The models were then applied without refitting to an RNA-only cohort of 43 donors with heart failure with preserved ejection fraction (HFpEF). Results: Across unseen cardiac cell types, CardioChrom increased area under the precision-recall curve by 0.0109 relative to within-cell-type RNA shuffling in 11 of 12 cell types and increased area under the receiver operating characteristic curve by 0.0179 in all 12. CardioChrom exceeded controls in both external cohorts and matched or exceeded the Transformer across 6 prespecified histone metrics. In HFpEF, 3,581 donor-level pathway-layer associations met global false discovery rate, sex-adjustment, and leave-one-donor-out robustness criteria. Endothelial cells showed prominent reconstructed histone-associated changes involving coagulation, inflammation, hypoxia, and profibrotic signaling. Von Willebrand factor expression and its reconstructed regulatory states were consistently reduced, and all 6 linked candidate cis-regulatory elements were classified as repressed. Motif-supported TF-cCRE-gene hypotheses were generated in endothelial cells and fibroblasts. Conclusions: CardioChrom extends RNA-only human heart failure cohorts toward cell-type-resolved virtual epigenomic and regulatory analysis while distinguishing predictions from direct measurements.

bioinformatics↗

bioRxiv2025

Lab-in-the-loop therapeutic antibody design with deep learning

Therapeutic antibody design is a complex multi-property optimization problem with substantial promise for improvement with the application of machine-learning methods. Towards realizing that promise, we introduce "Lab-in-the-loop," a new approach that orchestrates state-of-the-art repertoire mining methods, generative machine learning models, multi-task property predictors, active learning ranking and selection, and in vitro experimentation in a semi-autonomous, iterative optimization loop. By automating the design of antibody variants, property prediction, ranking and selection of designs to assay in the lab, and ingestion of in vitro data, we enable an end-to-end approach to developing computationally-informed therapeutic antibody design pipelines. We apply lab-in-the-loop to eleven seed antibodies obtained via animal immunization with four clinically relevant antigen targets: EGFR, IL-6, HER2, and OSM. Over 1,800 unique antibody variants are tested throughout four rounds of iterative optimization identifying 3-100x better binding variants for all targets and 10/11 seeds, with the best binders exceeding 100 pM affinity, demonstrating a process by which end-to-end machine learning can be developed for therapeutic antibody development.

bioengineering↗