Showed specific AML patients have genome-wide defects in their DNA methylation profiles, this could be exploited for theraphy/diagnosis.
We have identified promoter features of genes under long range gene regulation using statistics and bioinformatics methods.
We showed high-occupancy target regions associated with ChIP-seq noise using ML and stats based methods.
Deep learning method for integrating multi-omics data from cancer genomics. Applied on colorectal cancer for refining subtypes.
Integrate genomic data sets via meta-region plots and heatmaps. Works with BAM, BigWig, Bed files
Analyze methylation data from high-throughput BS-seq experiments
We showed WT1 mutation causes similar epigenome defects to IDH1/2 and TET2 mutations in leukemia