Yang Liu earns PhD on reference bias in disease genomics

About

Yang Liu has successfully defended her PhD thesis, Reference Bias and Variant Interpretation in Human Disease Genomics, at KAUST, under the supervision of Professor Robert Hoehndorf. Her work examines a foundational assumption in clinical genomics: that pathogenic variants are reliably observable given existing reference genomes.

The first half of the thesis focuses on disease-driven variant interpretation under complex biological contexts. Using hepatocellular carcinoma as a case study, Yang shows how somatic and germline variants can be integrated with functional and biological knowledge to elucidate disease mechanisms in heterogeneous, high-noise settings. A second study on retinoblastoma, a genetically well-defined rare disease, extends this by exploring how variant prioritisation strategies can be formalised into transferable frameworks that generalise beyond a single disease context.

The second half challenges a deeper assumption: that the linear reference genome is a neutral substrate for variant discovery. Focusing on structural variants and population-scale rare disease cohorts, Yang demonstrates that reliance on linear references introduces systematic reference bias, particularly for population-specific and structurally complex variation. By integrating population-specific pangenome references with short-read sequencing workflows, the thesis shows that the choice of reference model fundamentally alters variant discovery, representation, and downstream interpretability, with direct consequences for diagnostic yield.

Collectively, Yang's work makes the case for representation-aware genomics, where the methodological choices used to observe variants are treated as first-class scientific objects rather than implicit assumptions. Congratulations, Dr. Liu!