Understanding the sequence coverage distribution results

The results are a composite dataset, containing several files. Click on the eye icon in the history panel to show a page with some information about the run, along with links to the actual results: in this case a PDF graph and a text table. The graph gives a good overall view of the coverage, but when there are many individuals it can be difficult to distinguish which line is for who. The table provides easier access to evaluate how many SNPs were found at each coverage level in each individual.

Each column of the table represents a coverage level (read count), and each row corresponds to a single individual. The entries indicate what percentage of the SNPs were observed at or below that coverage level in that individual. Our dataset has nine million SNPs but we only need about 100,000 for this analysis, so we can afford to lose well over ninety percent. Since each of the individuals has a different total number of reads [I don't follow; these are percentages of the 9.4M SNPs, not of the number of reads, right?], we don't know exactly how many SNPs we'll get, but a coverage cutoff of "8" looks like a reasonable starting point (blue box). If the resulting number of SNPs is not what we expected, we can always come back and choose a different threshold.