BALTIMORE (GenomeWeb) – At the American Society of Human Genetics meeting here this week, Washington University in St. Louis' McDonnell Genome Institute cancer genomics researcher Li Ding touched on the somatic mutation and clonal evolution patterns that can be gleaned from sequencing tumor, matched normal, and relapse tumor samples at far greater depths than is the current standard.
As part of a broader discussion on sequencing breadth, depth, analyses, and big data challenges associated with cancer genomics, Ding pointed to an acute myeloid leukemia case that she and her colleagues sequenced to coverage depths exceeding 300-fold, on average, followed by a validation sequencing and targeting sequencing done at depths of between 1,000-fold and 250,000-fold coverage.
The case was described in a paper published in Cell Systems last month that was led by WUSTL researcher and McDonnell Genome Institute Director Richard Wilson. For that study, the researchers also evaluated several different sequencing and analysis protocols — from strategies used to produce sequencing libraries to algorithms used to align the resulting sequence reads and call variants.
For their assessment of ultra-deep sequencing, Ding explained, the researchers focused on samples from a patient known as AML31, a man who was diagnosed with AML when he was 55 years old and relapsed 16 months after treatment and remission.
His tumor, normal, and relapse samples were originally sequenced to average depths of 24-fold coverage, 32-fold coverage, and 38-fold coverage, respectively, for a paper on clonal evolution published in Nature in 2012.
To investigate the consequences of delving deeper with genome coverage, the team used the Illumina HiSeq instrument to resequence the primary tumor to 312-fold coverage and a matched normal skin sample to 121-fold coverage, on average.
The exomes of tumor, normal, and relapse samples were also sequenced to average depths ranging from 250-fold to more than 430-fold coverage, and deep RNA sequencing was done on the primary tumor and matched normal samples.
After uncovering an initial set of suspected mutations, the researchers did further validation sequencing with a combination of custom capture reagents, Illumina sequencing, and Ion Torrent sequencing, producing up to 250,000-fold targeted coverage in some cases.
Compared with the 118 somatic mutations previously detected in the AML founding clone, Ding explained, the deep sequencing coverage made it possible to narrow in on more than 1,300 verified somatic mutations. That set included nine potential cancer drivers, among them, a TP53 mutation that expanded to become present in nearly one-fifth of the relapse sample.
"Regardless of which sequencing approach is used, it is clear that given the propensity of rare subclones to harbor mutations that contribute to therapy resistance, there is an urgent need to become more adept at discovering low-frequency events at presentation," Ding, Wilson, and their co-authors wrote in the Cell Systems study.
"This requires pushing the boundaries of today's sequencing platforms, developing a better understanding of technical details like error rates and lower bounds of detection, and developing new sets of best practices," they added.
These and other findings have already altered the team's view of the clonal complexity at play over time in the AML case, Ding explained, pointing to the importance of going as deep as possible when trying to understand tumor biology, potential treatment targets, and the trajectory of relapse.
She noted that co-author Timothy Ley, also with the WUSTL McDonnell Genome Institute, collected samples from the AML31 individual at time points between diagnosis and relapse that are expected to contain genomic clues that may further refine the team's view of the clonal architecture behind the case.