Convex approaches to isolate the shared and distinct genetic components of complex traits.

Convex optimization with Clorinn accurately isolates shared and unique genetic components across complex traits, recovering 200 latent factors from GWAS data.

• Why it matters: Understanding shared genetic influences among diseases can reveal biological mechanisms and improve disease classification, but current methods struggle with noise and contamination.
• What they did: The authors developed Clorinn, a Python library that models genetic summary statistics as a sum of low-rank and sparse components, using convex optimization to ensure reproducibility and robustness.
• The result: Clorinn successfully recovers disease-group structures and latent factors from large GWAS datasets, enabling more reliable analysis of complex genetic architectures and their shared biological processes.