Discovery of sparse, reliable omic biomarkers with Stabl
Inventors & their inventions
Axial: https://linktr.ee/axialxyz
Axial partners with great founders and inventors. We invest in early-stage life sciences companies such as Appia Bio, Seranova Bio, Delix Therapeutics, Simcha Therapeutics, among others often when they are no more than an idea. We are fanatical about helping the rare inventor who is compelled to build their own enduring business. If you or someone you know has a great idea or company in life sciences, Axial would be excited to get to know you and possibly invest in your vision and company. We are excited to be in business with you — email us at info@axialvc.com
The paper introduces Stabl, a new machine learning method for identifying sparse and reliable biomarker candidates from high-dimensional omics data.
Stabl combines subsampling, noise injection via knockoffs or permutations, and minimization of a false discovery proportion (FDP) surrogate to select features in a data-driven way. Benchmarking on synthetic and real datasets shows Stabl selects fewer features than Lasso, Elastic Net, etc. while maintaining predictive performance. This improves model interpretability. Stabl is flexible and modular, allowing different sparse regression methods and noise injection techniques to be used based on the dataset characteristics.
Case studies demonstrate Stabl's ability to extract sparse biological signatures predictive of outcomes like preeclampsia, COVID-19 severity, time to labor, and surgical site infections. Key advantages of Stabl are its data-driven threshold optimization, improved reliability, retained predictivity, and ability to integrate multi-omics data for biomarker discovery.
Overall, Stabl shows promise as a machine learning pipeline to facilitate clinical translation of omics studies by honing in on robust candidate biomarkers within predictive modeling.
https://www.nature.com/articles/s41587-023-02033-x
