Shayan Nasiriboukani
Academic and research departments
School of Computer Science and Electronic Engineering, Centre for Vision, Speech and Signal Processing (CVSSP).About
My research project
LOMsGWAS: Large Omics Models for Multi-omics and Genome Wide Association StudiesThe research proposal will focus on developing explainable, interpretable, and highly efficient LOMs for multi-omics applications, including GWAS, radiomics, transcriptomics, metabolomics, and other omics-based applications. The research will leverage SSL to train LOMs, beginning with radiomics.
It will then focus on integrating other omics modalities to enhance disease understanding. This will enable tasks such as medical Q&A, report generation, grounding, medical reasoning, GWAS, predicting transcription factor binding sites, splice sites, and identifying promoter regions. The LOM will incorporate a bespoke deep neural network (DNN) as a tokenizer, significantly reducing compute and memory requirements while maintaining the explainability and interpretability of LOMs.
Supervisors
The research proposal will focus on developing explainable, interpretable, and highly efficient LOMs for multi-omics applications, including GWAS, radiomics, transcriptomics, metabolomics, and other omics-based applications. The research will leverage SSL to train LOMs, beginning with radiomics.
It will then focus on integrating other omics modalities to enhance disease understanding. This will enable tasks such as medical Q&A, report generation, grounding, medical reasoning, GWAS, predicting transcription factor binding sites, splice sites, and identifying promoter regions. The LOM will incorporate a bespoke deep neural network (DNN) as a tokenizer, significantly reducing compute and memory requirements while maintaining the explainability and interpretability of LOMs.