Because data from one large study of autism provided genomic coordinates for an exceptionally small fraction of CNVs (31 CNVs in 1562 samples) found in affected individuals, we have repeated the analysis summarized in Table 2 from our recent paper excluding data from that particular study. The new analysis includes CNVs from a total of 10,972 individuals: 5,674 controls as before, but only 5,298 affected individuals as compared to 6,860 in our previous analysis. The ranking among the top-scoring loci has changed slightly (see Alternate Table 2 below), and p values have uniformly decreased (suggesting stronger associations) with two exceptions: duplications at the VCFS locus went from p = 0.330 to 0.454 and duplications overlapping the Terminal 22 del syndrome region went from p = 0.160 to 0.167. The qualitative conclusions we made on the basis of the previous analysis and other evidence (that 3q29,16pl2, and 15q25 merit further investigation) have not changed. We believe the new analysis more accurately reflects CNV frequencies in individuals with neurologic disease
展开▼