Table 1. Variant summary.

Variant Type			Call Rate Cutoff	Mean Error Rate	Total Variants	Variants Not Passing QC	Variants Passing QC
Non-singletons	RS	SNVs	0.9	0.12%	6,879,488	1,332,939	5,546,549
		Insertions	0.9	0.14%	363,973	259,472	104,501
		Deletions	0.9	0.13%	352,563	222,185	130,378
	Novel	SNVs	0.99	0.68%	1,330,086	667,306	662,780
		Insertions	0.99	0.97%	577,508	547,690	29,818
		Deletions	0.99	0.42%	306,776	232,648	74,128
Singletons	RS	SNVs	0.99	0.42%	277,650	89,448	188,202
		Insertions	0.99	1.37%	28,185	27,089	1,096
		Deletions	0.99	0.50%	24,822	21,613	3,209
	Novel	SNVs	0.99	0.14%	5,522,109	5,302,013	220,096
		Insertions	0.9	1.99%	1,756,253	1,728,591	27,662
		Deletions	0.99	1.50%	756,632	736,385	20,247
Total		SNVs	-	0.27%	14,009,333	7,391,706	6,617,627
		Insertions	-	1.10%	2,725,919	2,562,842	163,077
		Deletions	-	0.51%	1,440,793	1,212,831	227,962
		All Variants	-	0.37%	18,176,045	11,167,379	7,008,666

Variants present in 98 Hutterite genome sequences, shown by singleton (one alternative allele among the 98 sequenced samples) vs. non-singletons, by novelty (variants with and without rs numbers in dbSNP135), and by variant type (SNVs, insertions, deletions). Insertions and deletions are called relative to the NCBI build 37 reference human genome. The mean genotyping calling error rate is estimated using IBD2 segments (see text). QC is based on a combination of using CGI high-quality genotype calls and the shown call rate cutoffs.