Sensitive and specific detection of mosaic chromosomal abnormalities using the Parent-of-Origin-based Detection (POD) method

Table 4 Analysis of large low-level mosaic abnormalities in the AGRE autism dataset

Region	Sample	Chr	Size(SNPs)	% Mosaic	triPOD	PSCN	MAD	BAFseg	genoCNA
1	AU031003	5	9664	2.1 - 6.4	0.96	0.02	0	0.02	0.06
2	AU036104	22	4970	4.9 - 12.6	0.89	0.02	0	0	0
3	AU051503	7	7707	3.7 - 10.7	0.69	0.03	0	0	0
4	AU068604	5	1499	5.6 - 16.5	1	0	0	0	0
5	AU072004	1	19981	6.5 - 18.7	1	0.93	0	0.03	0.05
6	AU073006	8	1710	8.4 - 24	0.97	0.87	0.99	0.05	0
7	AU0871303	11	5219	2.8 - 8.1	0.99	0	0	0.02	0.03
8	AU1271304	9	3202	2.5 – 6.8	0.73	0.39	0	0.01	0.02
9	AU1285302	13	7455	7.3 – 21.5	1	1	0.93	1	0.91
10	AU1346302	19	3241	6.7 - 19.2	1	0.45	0	0	0
11	AU1462303	9	15601	2.2 - 6.0	0.81	0.48	0	0	0
12	AU1585303	9	15421	2.8 - 8.1	0.92	0.02	0	0.01	0.07

The sensitivity of abnormal region detection for 12 samples in the AGRE autism dataset harboring a large low-level mosaic abnormality is presented. The samples were analyzed by triPOD, BAFsegmentation (unpaired), genoCNA, MAD, PennCNV joint, and PSCN. (Sensitivity results < 0.01 are reported as 0). The % Mosaic column is the estimated percent mosaicism calculated, using BAF values >= the baseline median and < 0.7, as the ratio of the distance of the abnormal BAF median from the baseline median and the expected distance of a 100% mosaic abnormality from the baseline median. For low-level mosaicism, the type of abnormality is difficult to distinguish based on LRR values, thus a range is provided to account for regions resulting from a deletion or UPD event (smaller percentage) or an amplification.

ISSN: 1471-2164