Figure 2

The Fmrp targets are long transcripts, and long transcripts overlap with autism candidate genes. (A) Histogram of coding sequence (Cds) lengths of Fmrp targets (purple) compared to an equal length list of random brain-expressed genes (grey). Fmrp targets are significantly longer than random brain-expressed genes (P < 2.2e − 16, t test). (B) Cds lengths correlate moderately with Fmrp HITS-CLIP counts (0.48). (C) A sample of random genes selected to match the expression of the Fmrp targets also significantly overlap with the rDNV and SFARIdb. (D) Histogram of the P values resulting from 1,000 Fisher’s exact tests using random gene sets, each sampled to match the Fmrp target’s Cds length, compared to SFARIdb (top panel) or rDNV genes (bottom panel). Sampled sets were generally less significant than the true Fmrp list (red arrow, P < 1.2e − 11, P < 6.1e − 12), but most were more than nominally significantly (blue line, P < 0.05).