Evaluation of whole genome sequencing utility in identifying driver alterations in cancer genome

Abstract

In cancer genome analysis, identifying pathogenic alterations and assessing their effects on oncogenic processes is important. Although whole exome sequencing (WES) can effectively detect such changes, driver alterations could not be identified in 27.8% of the cases, according to a previous study. The objectives of the present study were to evaluate the utility of whole genome sequencing (WGS) and clarify its differences with WES in terms of driver alteration detection. For this purpose, WGS analysis was conducted on 177 driverless WES samples, selected from 5,480 fresh frozen samples derived from 5,140 Japanese patients with cancer. These samples were selected as primary tumor, both WES and transcriptome profiling were performed, estimated tumor content of ≥ 30%, and no driver alterations were identified by WES. WGS identified driver and likely driver alterations in 68.4 and 22.6% of the samples, respectively. The most frequent alteration type was oncogene amplification, followed by tumor suppressor gene deletion and small variants located outside the coding region. In the remaining 9.0% of samples, no such signals were identified; therefore, further investigations are required. The current study clearly demonstrated the role and utility of WGS in identifying genomic alterations that contribute to tumorigenesis.

Keywords: Cancer genome analysis; Driver alteration; Gene expression profiling; Whole genome sequencing.

Fig. 1

Selection of driverless samples based on whole exome sequencing and transcriptome analysis. (a) Flowchart of the sample selection procedure. A total of 177 samples were selected from 5,480 samples through 5 steps. “T” and “N” represent tumor and normal, respectively. (b) Distribution of cancer type in 177 samples. Cancer type was sorted in descending order based on the number of samples in that cancer type. Abbreviations of cancer types are shown in Table S1. (c) Distribution of cancer types in four different samples according to the selection procedure shown in (a). The numbers in square brackets correspond to those shown in (a).

Fig. 2

Frequency of driver alterations identified by whole genome sequencing and transcriptome analysis. (a) The percentage of samples with driver, likely driver, and likely driver supported by gene expression alterations only are shown. The numbers in brackets under cancer type represent the sample size. (b) Relative frequency of five types of genomic alterations. The numbers in brackets under cancer type show the number of total alterations in that cancer type. “g” stands for germline.

Fig. 3

Frequency of genes with driver alterations. The mean number of driver alterations per sample is shown as the frequency on the y-axis. Genes were categorized by function and sorted by function and gene names. “Epi. reg.” and “Prot. hom.” represent epigenetic regulation and protein homeostasis, respectively. “g” stands for germline.