Title:	Isoform discovery with controlled FDR
DNr:	NAISS 2026/4-220
Project Type:	NAISS Small
Principal Investigator:	Mattias Hansen <mattias.hansen@ki.se>
Affiliation:	Karolinska Institutet
Duration:	2026-02-04 – 2026-05-01
Classification:	10203
Keywords:

Abstract

Public DIA proteomic datasets contain substantial unexplored signal for novel protein isoforms and non-canonical translation products, but discovery at scale is limited by the size of predicted spectral libraries and the difficulty of maintaining robust false discovery control across heterogeneous hypothesis spaces. We will build and evaluate long-read–informed proteogenomic DIA workflows to improve discovery power while keeping error rates well-calibrated for predicted and previously unannotated peptides. The outputs will include reproducible analysis results across multiple public cohorts and a validated set of best-practice library design and filtering strategies that can be reused by other researchers. The work is CPU-intensive because it requires repeated DIA searches across multiple library configurations and validation replicates.