Recent studies have used whole-genome sequencing to estimate subclonal populations in tumours and have linked this heterogeneity to clinical outcomes. Many algorithms have been developed to perform subclonal reconstruction but their variability and consistency is largely unknown. To quantify this effect, we evaluated six pipelines for reconstructing the evolutionary histories of 293 localized prostate cancers from single samples and 10 from multi-region sampling to probe the heterogeneity of subclonal reconstruction. We identified extensive variance across pipelines in the subclonal architectures they predict and in their assignments of CNAs and SNVs to different parts of the evolutionary tree. Individual pipelines showed consistent types of bias, with those using SomaticSniper and Battenberg preferentially predicting homogenous cancer cell populations while those using MuTect tending to predict multiple populations of cancer cells. Subclonal reconstructions using multi-region sampling showed that single-sample reconstructions systematically underestimated intra-tumoural heterogeneity, detecting on average fewer than half of the cancer cell populations identified by multi-region sequencing. These biases suggest caution in interpreting the specific architectures and subclonal variants identified, particularly from single-sample reconstructions.*   CNAs    :   Copy Number Aberrations    SNVs    :   Single Nucleotide Variants    TSP    :   phylogenetic reconstruction pipeline TITAN-SomaticSniper-PhyloWGS    TMP    :   phylogenetic reconstruction pipeline TITAN-MuTect-PhyloWGS    BSP    :   phylogenetic reconstruction pipeline Battenberg-SomaticSniper-PhyloWGS    BMP    :   phylogenetic reconstruction pipeline Battenberg-MuTect-PhyloWGS

Quantifying the Influence of Mutation Detection on Tumour Subclonal Reconstruction

Scan to connect with one of our mobile apps

Coinbase Wallet app

Coinbase app

Or try the Coinbase Wallet browser extension