Abstract
Background DNA sequencing of archived dried blood spots (DBS) collected by newborn screening programs constitutes a potential health resource to study newborn disorders and understand genotype-phenotype relationships. However, its essential to verify that sequencing reads from DBS derived DNA are suitable for variant discovery.
Results We explored 16 metrics to comprehensively assess the quality of sequencing reads from 180 DBS and 35 whole blood (WB) samples. These metrics were used to assess a) mapping of reads to the reference genome, b) degree of DNA damage, and c) variant calling. Reads from both sets mapped with similar efficiencies, had similar overall DNA damage rates, measured by the mismatch rate with the reference genome, and produced variant calls sets with similar Transition-Transversion ratios. While evaluating single nucleotide changes that may have arisen from DNA damage, we observed that the A>T and T>A changes were more frequent in DNA from DBS than from WB. However, this did not affect the accuracy of variant calling, with DBS samples yielding a comparable count of high quality SNVs and indels in samples with at least 50x coverage.
Conclusions Overall, DBS DNA provided exome sequencing data of sufficient quality for clinical interpretation.
Competing Interest Statement
Uma Sunderam and Rajgopal Srinivasan are employees of TCS. Aashish Adhikari is an employee of Illumina, Inc. Kunal Kundu was an employee of TCS. Steven E. Brenner receives support at the University of California Berkeley from a research agreement from TCS. Jennifer Puck's spouse is employed at Invitae, a clinical DNA sequencing company.
Footnotes
↵* email: uma.sunderam{at}tcs.com, rajgopal.srinivasan{at}tcs.com
updated ORCID for two authors