Copy-number variations (CNVs) are often detected when alterations in sequencing depth at a given locus indicate gain or loss of genetic material. Researchers commonly use exome sequencing to identify genetic mutations, but it is especially difficult to apply this approach to CNV detection because of inherent difficulties in determining sequencing depth from exome-sequencing data. The confounding factors primarily stem from the noncontiguous nature of exome sequencing and differences in the sequence itself. Accordingly, researchers must derive normalizing factors or turn to alternative techniques such as array-comparative genomic hybridization to gather high-quality data about CNVs. By leveraging principal-component analysis to normalize read depth and following this with Hidden Markov Model analysis, Fromer et al. developed eXome Hidden Markov Model (XHMM), a method that identifies CNVs from exome-sequencing data with high confidence. Because XHMM works best with larger data sets, it is well suited for whole-genome evaluation, produces scores from which CNV size can be estimated, and can statistically genotype a CNV in a given population. In evaluations of 90 schizophrenia trios, high-quality scores for CNVs and few Mendelian violations indicated that the method works robustly. Furthermore, the use of XHMM allowed the authors to identify an increased burden of partial gene disruptions in schizophrenia-affected individuals in comparison to normal controls. Armed with XHMM, researchers can use exome sequencing to identify an expanding assortment of genetic mutations and to perform genotyping with increasing efficiency.
展开▼
