arXiv:2512.17678v2 Announce Type: replace Abstract: Selecting compact and informative gene subsets from single-cell transcriptomic data is essential for biomarker discovery, improving interpretability, and cost-effective profiling. However, most existing feature selection approaches either operate as multi-stage pipelines or rely on post hoc feature attribution, making selection and prediction weakly coupled. In this work, we present YOTO (you only train once), an end-to-end framework that jointly identifies discrete gene subsets and performs prediction within a single differentiable architect
The proliferation of complex biological datasets, particularly single-cell transcriptomics, necessitates more efficient and interpretable methods for biomarker discovery and data analysis, making joint optimization critical.
This work directly addresses a key bottleneck in omics data analysis, enabling more accurate, interpretable, and cost-effective selection of relevant biological features, accelerating drug discovery and personalized medicine.
Traditional multi-stage or post hoc feature selection in omics is replaced by an end-to-end differentiable framework, YOTO, leading to better-coupled selection and prediction outcomes.
- · Biotech companies
- · Pharmaceutical research
- · AI in healthcare developers
- · Personalized medicine initiatives
- · Traditional bioinformatics pipelines reliant on sequential optimization
Faster and more accurate identification of disease biomarkers and therapeutic targets.
Reduced R&D costs and accelerated time-to-market for new drugs and diagnostics.
Enhanced ability to develop highly personalized treatments based on individual molecular profiles, leading to more effective therapies and better patient outcomes.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG