Rare copy number variant analysis in case–control studies using snp array data: a scalable and automated data analysis pipeline

Table 1 Calling and QC sub-pipeline report

Module 1: Data Conversion
Generate signal intensity file	700,079 markers 6,112 samples

Module 2: Data Calling
	Initial samples	Final samples	Lost	Initial calls	Final calls	Lost
Raw data	6,112	–	–	98,702	–	–
Module 3: Data Clean
Filters	Initial samples	Final samples	Lost	Initial calls	Final calls	Lost
Default parameters*	6,112	6,012	100	98,702	71,010	27,692
Clean Immunoglobulin regions	6,012	6,012	–	71,010	70,436	574
Clean centromere and telomere regions	6,012	6,012	–	70,436	63,202	7,234
Merging calls**	6,012	6,012	–	63,202	60,705	2,497

Module 3: Data Clean
Filters	Initial samples	Final samples	Lost	Initial calls	Final calls	Lost
Default parameters*	6,112	6,012	100	98,702	71,010	27,692
Clean Immunoglobulin regions	6,012	6,012	–	71,010	70,436	574
Clean centromere and telomere regions	6,012	6,012	–	70,436	63,202	7,234
Merging calls**	6,012	6,012	–	63,202	60,705	2,497

The table summarizes the samples included and excluded at each module in the calling and QC sub–pipeline. Final samples and calls, after QC, are in bold
^*LRR_SD < 0.3, BAF_drift < 0.01, |WF|< 0.05, NumCNV > 50
^**fraction: 0.5 and 0.4. In this step, calls were not lost, but the number decreased because two or more calls can be combined into a unique

ISSN: 1471-2105