REViewer reports various summary measurements, called quality metrics, that describe key properties of read pileups. Quality metrics make it possible to automate assessment of large collections of STR genotype calls either (a) by selecting a series of thresholds to stratify the value of each metric into "good", "suspicious", and "bad" categories, or (b) by using more flexible statistical / machine learning approaches.
This document describes the quality metrics reported by REViewer. All quality
metrics are stored in a tab-separated (TSV) file <output prefix>.metrics.tsv.
If you have suggestions for additional metrics, please consider creating an
issue.
In general, sequencing depth of a whole-genome sequencing sample is the average number of reads that overlap any given genomic position in that sample. (For example, a depth of 30x means that a base would be overlapped by 30 reads on average.) The allele depth metric is an extension of this concept to STRs: It reports the sequencing depth of each STR allele. The diagram below shows an example of a well-genotyped repeat (left) where both STR alleles have the expected sequencing depth and an example (right) where the size of the long allele may be overestimated.
