Fig. 7

Lack of CpG islands and restricted expression is a feature of tandemly arrayed AT-rich genes. A Promoter CpG island frequency (from [66]) in isochores of varying gene content. Same plots as in Fig. 2E, recolored to highlight promoter characteristics. Most common gene name prefixes in the isochore are labeled. B Multigene score versus isochore GC% for isochores containing tandem duplicated genes. Color represents fraction of genes in the isochore with promoter CpG islands. C–E Distribution of promoter CpG rate for genes binned by C isochore GC% decile, D k-means cluster assignment, or E tandem duplication score. We counted CpGs from −750 to +250 bp around the gene TSS and compared to the expected rate for a sequence of that GC content. F–H Distribution of tissue-level gene expression for genes binned by F isochore GC% decile, G their k-means cluster assignment, or H tandem duplication score. GTEx data from 54 human tissues was binarized to “expression” or “no expression” based on RPKM of 5. AT-rich genes and those in paralogous arrays are detected in few or no sampled GTEx tissues, while GC-rich genes are frequently detected in all sampled tissues. Black bars depict medians