Theses Doctoral

Integration of Germline and Somatic Variation in Tumor Data

Dewal, Ninad Pradeep

During tumor inception and progression, culprit gene variants confer selective advantage to progenitor cancer cells, allowing them to outcompete normal cells and proliferate uncontrollably. Both regions of somatic amplification as well as germline DNA sequence changes may be variants that are positively selected by the tumor. Traditionally, these two variant classes have been studied independently. While many discoveries have been made in such a manner, independent examination of these classes possesses certain limitations. Integrated examination of these two classes holds the potential to reveal specific nucleotide alleles that are amplified in the tumor, which in turn may reveal proximal genes. We present methods that focus on such integration. The first, the Amplification Distortion Test (ADT), aims to detect nucleotide alleles that are selectively amplified across tumor samples. Motivated to apply ADT on nascent next generation sequencing data, we developed a novel Hidden Markov Model-based method - Haplotype Amplification in Tumor Sequences (HATS) - that analyzes tumor and matched normal sequence data, along with training data for linkage information, to infer amplified alleles and haplotypes in regions of copy number gain. HATS is designed to handle biases in read data as well as accommodate rare variants. We demonstrate that HATS infers the amplified alleles more accurately on simulated and real tumor data than does an alternate naïve approach, especially at low to intermediate sequence coverage levels, and when allele-specific biases or stromal contamination is present. We present these methods with the motivation that they may aid the cancer community in identifying novel causal or associated putative variants.


  • thumnail for Dewal_columbia_0054D_10422.pdf Dewal_columbia_0054D_10422.pdf application/pdf 7.77 MB Download File

More About This Work

Academic Units
Biomedical Informatics
Thesis Advisors
Pe'er, Itschack G.
Ph.D., Columbia University
Published Here
November 6, 2017