Huimin Li

This paper benchmarks the Evo2 genomic AI model in cancer genomics, showing FP8 precision on H800 GPUs improves inference speed and accuracy, while larger models perform worse on non-FP8 hardware.

This study identifies five copy number signatures in prostate cancer using sigminer, revealing distinct etiologies and improved clinical outcome associations compared to single base substitution signatures.

This paper critiques a previous study that claimed no neoantigen depletion signal in untreated cancer genomes, arguing that the study's methodology is flawed due to misalignment between defined HLA-binding regions and actual immunogenic neoantigens.