RESEARCH SUMMARY
My research focuses on developing algorithmic methods for managing, storing, communicating and analyzing high-throughput cancer sequencing data.

Research
A long-time focus of my lab has been discovery and interpretation of large-scale (especially structural) genomic and transcriptomic alterations in tumor samples. Our algorithmic methods for genomic structural variation discovery, including VariationHunter, CommonLAW, DeStruct and NovelSeq, were the first with the ability to handle novel insertions, deletions, inversions and duplications in repetitive regions of the human genome. More recently I have been interested in applying our algorithmic techniques to exact genotyping of highly repetitive, structurally variant genes, e.g., those involved in drug metabolism – for which my group has developed Cypiripi and Aldy methods, and immunoglobulin heavy chain region, for which my group has developed Immunotyper. My group has also contributed to the identification and quantification of transcriptomic aberrations, in particular gene fusions, as well as genic inversions, duplications and deletions in cancer samples. Leading computational methods we have developed include DeFuse, NFuse, Comrad MiStrVar and SVICT (which handles circulating cell-free tumor DNA data). Our most recent focus area is tumor heterogeneity and progression modeling, especially by the use of single-cell sequencing or multi-locus/time series sequencing (for which we have developed CITUP, CTP-Single, Remix-T, BSCITE, PhISCS, CONETT, PhISCS-BnB); my lab has developed the first deep learning method for inferring the progression tree of a tumor. We also work on network-aided, integrative analysis of genomic and transcriptomic sequence data from tumor samples (Hit’nDrive and cdCAP). We have several additional interests within "algorithmic biology" including (i) mapping and variant calling (of/via especially reads from repetitive regions of the genome – or involving reads with high error rates – examples include mrFAST, mrsFAST, drFAST and lordFAST, or reads extracted from cell free tumor DNA - e.g. SINVICT), (ii) genomic data compression (SCALCE, DeeZ and AssemblTrie), (iii) secure/privacy preserving computing (PrivStrat, SkSES and SMac) and (iv) metagenomic binning (CAMMiQ).