Modeling the DNA copy number aberration patterns in observational high-throughput cancer data

Area of Science:

• Genomics
• Cancer Biology
• Computational Biology

Background:

• Genomic aberrations accumulate over time, reflecting cancer progression.
• High-resolution technologies like array Comparative Genomic Hybridization (aCGH) and Massive Parallel Sequencing (MPS) measure DNA copy number aberrations.
• Time is a latent variable in cancer progression, often proxied by the number of aberrations.

Purpose of the Study:

• To propose a continuous time Markov chain model for genomic aberration accumulation.
• To estimate model parameters from high-dimensional DNA copy number data.
• To obtain posterior time estimates for cancer advancement and predict aberration locations.

Main Methods:

• A continuous time Markov chain model is developed.
• Model parameters are estimated using penalized maximum pseudo-likelihood, penalized maximum likelihood, and method of moments.
• Time is integrated out to derive the distribution of observed DNA copy number data.

Main Results:

• The proposed model allows for the estimation of cancer progression time using DNA copy number data.
• Posterior time estimates and likely aberration locations are obtained for individual cancer samples.
• A simulation study compares the performance of three estimation methods.

Conclusions:

• The methodology provides a framework for analyzing cancer progression dynamics using genomic aberration data.
• The model can be applied to real cancer datasets for insights into disease advancement.
• Accurate estimation of cancer advancement and aberration locations is achievable.