My Clinical appointment is as an instructor working at the intersection of Molecular Genetics (interpretation of genetic tests ordered by doctors for patients; 2023-4) and Clinical Informatics (use of information systems and clinical data mining to improve healthcare; 2023-5) in Laboratory Medicine at Michigan.
Scientific interests flow from the belief that in ~4 years time today's genomics (short-read sequencing aligned to a single linear reference genome, and presented as SNVs in .vcf file, without epigenetic information) will be entirely supplanted by emerging technologies that capture long, contiguous stretches of genetic and epigenetic variation. Such a rapid transformation entails development of data structures and software that can represent diverse, gapless, phased, human haplotypes and their annotations.
Private ventures formerly related to meta-analytics or "bioinformatics" - a field that will be wholly subsumed into AI research. As such, presently, interests have shifted towards:
1) neurosymbolic AI design, integration, and deployment
2) (genome) foundation model design, training, and deployment
3) EMR-molecular genetic-billing interoperability (to improve patient care, but also improved billing, e.g. value based care, becomes possible)
4) application of multimodal ai to clinicogenomic data
5) complete and total interoperability between all genomic nomenclatures without loss of information (ask me)
6) last and most importantly, i am a technical and political advocate for pangenomics; GRCh38 must be phased out as quickly as possible or our patients and our scientific efforts pay the price
