Dark matter RNA illuminates the puzzle of genome-wide association studies

Abstract

In the past decade, numerous studies have made connections between sequence variants in human genomes and predisposition to complex diseases. However, most of these variants lie outside of the charted regions of the human genome whose function we understand; that is, the sequences that encode proteins. Consequently, the general concept of a mechanism that translates these variants into predisposition to diseases has been lacking, potentially calling into question the validity of these studies. Here we make a connection between the growing class of apparently functional RNAs that do not encode proteins and whose function we do not yet understand (the so-called 'dark matter' RNAs) and the disease-associated variants. We review advances made in a different genomic mapping effort - unbiased profiling of all RNA transcribed from the human genome - and provide arguments that the disease-associated variants exert their effects via perturbation of regulatory properties of non-coding RNAs existing in mammalian cells.

Figure 1

Discovery of genome-wide association studies (GWAS) variants in different genomic elements and disease types. GWAS variants were assigned to a disease type (y axis) or non-disease traits. For each disease type, the P-value (x axis) for a skew towards a particular genomic category (x xis) was calculated (see Additional file 1: Supplementary Text). Numbers of unique GWAS variants for each of the genomic categories are shown as the purple bars; the corresponding numbers for each of the disease types are shown in parenthesis. Only disease types with >100 GWAS variants are shown. CDS, coding DNA sequence (coding regions of known genes); UTR, untranslated region (non-coding regions of known genes; promoter and intronic regions are those of known genes). See Additional file 1: Supplementary Text for details of the analysis.

Figure 2

A genomic view of the 8q24 region upstream of the MYC gene. For details, see text.