Fmrp targets or not: long, highly brain-expressed genes tend to be implicated in autism and brain disorders

Many studies have demonstrated a robust statistical overlap between genes whose transcripts are reported as Fragile X Mental Retardation Protein (Fmrp)-binding targets and genes implicated in various psychiatric disorders, including autism. However, it is not clear how to interpret this overlap as the Fmrp protein itself is not considered to be central to all instances of these conditions.

We tested whether Fmrp binding may be a proxy for some other features of these transcripts. Reviewing recent literature on the cross-linking and immunoprecipitation (CLIP)-derived targets of Fmrp in the brain, and the literature on identifying genes thought to mediate autism and other psychiatric disorders, reveals that both appear to be disproportionately made up of highly brain-expressed genes. This suggests a parsimonious explanation-that the overlap between Fmrp targets and neuropsychiatric candidate genes might be secondary to simple features such as transcript length and robust expression in the brain. Indeed, reanalyzing Fmrp high-throughput sequencing of RNAs isolated by CLIP (HITS-CLIP) data suggests that approximately 60% of CLIP tag depth can be predicted by gene expression, coding sequence length, and transcript length. Furthermore, there is a statistically significant overlap between autism candidate genes and random samples of long, highly brain-expressed genes, whether they are Fmrp targets or not.

Comparison of known Fmrp-binding targets to candidate gene lists should be informed by both of these features.