Defining Negative Data in Tumor Heterogeneity Studies
MetadataShow full item record
Motivation: Assessing tumor heterogeneity on the molecular level involves comparison of somatic mutations between different tumor samples. One challenge lies in defining negative status, more specifically, when a mutation is not detected in a sample, whether the coverage is sufficient to claim it negative. Theoretically, there is no definitive way to determine a coverage threshold as it depends on the relative abundance of the mutation. Most previous tumor heterogeneity studies used a universal coverage cutoff when defining negative status while ignoring the difference in the relative abundance of variants. The result could be misclassification between negative and unknown statuses. Results: We hypothesized that defining negative status on individual variant’s basis with relative abundance information is more accurate than using a universal coverage cutoff. Using a unique dual-platform sequencing data set, our method produced better concordance between the two platforms than the methods based on universal coverage cutoffs. Conclusion: We proposed a new method for any tumor heterogeneity study to define variant’s negative status with better flexibility and improved accuracy than using universal coverage cutoff.