Sample Variability Tracking in Population-Scale Studies as a Methodology for Reference Genome Characterization

The increase in the number of sequenced samples in whole-genome next generation sequencing studies provides a suitable framework to deeply characterize genomic regions, in terms of individual (or population) variability and error susceptibility. Mapping collected statistic scores and variant sites over reference genome after SNV/indel calling allows the evaluation of each region of interest, and consequently, the establishment of an a priori confidence interval that could explain most of the false positive results in posterior downstream analysis, such as, genome-wide association studies. We used a window-based processing approach with a gaussian kernel that properly characterizes the neighborhood around a specific locus. This protocol provides a genome profile with an adjustable region size-level resolution. Our approach allows detecting intra/inter species highly conserved and polymorphic loci, as well as specific regions that are systematically uncovered or affected by noise, that could potentially conceal misassembled reference genome portions. We processed 141 citrus samples from CITRUSEQ/CITRUSGENN consortium which were aligned against the clementine genome. We detected 221 significant conserved regions between different species and 190 highly polymorphic sites (16,911,124bp and 18,119,035bp respectively). Moreover, we detected 179 poor quality regions (16,911,103bp) that consistently describe some conflictive loci in clementine genome. Finally, we constructed a database that collects some descriptive features for every base which will be available for the evaluation of specific biomarkers of future studies. In this study we demonstrated the validity of a valuable methodology for recent and future plant and animal genome studies.