Skip to main content
Fig. 7 | Genome Biology

Fig. 7

From: Characterizing the interplay between gene nucleotide composition bias and splicing

Fig. 7

Nucleotide composition bias and genome organization. a Correlation between the GC content of GC exons and AT exons, and the GC content of their hosting isochores defined by ISOFINDER. b Proportion of AT exons, GC exons, and control exons distributed across different isochore families. c Number of AT exons and GC exons present in individual isochores. Only isochores containing at least five GC exons or five AT exons are represented. The left and right panels represent isochores containing preferentially GC exons or AT exons, respectively. d Proportion of AT exons and GC exons in LADs annotated from three different datasets (1, fibroblasts; 2, resting Jurkat cells; 3, activated Jurkat cells) and in TADs annotated from three different cell lines (4, K562; 5, IMR90; 6, MCF7). e Correlation between the GC content of GC exons and AT exons, and the GC content of their hosting TADs, defined in the K562 cell line. f Number of AT exons and GC exons present in individual TADs annotated from the K562 cell line. Only TADs containing at least five GC exons or five AT exons are represented. The left and right panels represent TADs containing preferentially GC exons or AT exons, respectively. g GC-rich isochores and TADs contain a large number of genes (“gene core”) that are GC-rich and contain small introns. In contrast, AT-rich isochores, TADs, and LADs contain a small number of genes (“gene desert”) that are AT-rich and contain large introns. The regional nucleotide composition bias (over dozens of kbps) increases the probability of local nucleotide composition bias (e.g., at the gene and exon levels). Local nucleotide composition bias influences local chromatin organization at the DNA level (e.g., nucleosome density and positioning) as well as the splicing process at the RNA level. The high density of nucleosomes and GC nucleotides (upper panel) could generate a “smooth” transcription across small genes, favoring synchronization between transcription and splicing. The high density of GC nucleotides increases the probability of secondary structures at the 5′ ss, with consequences on splicing recognition during the splicing process. This constraint could be alleviated by splicing factor (SF; in blue) binding to GC-rich sequences, which enhances U1 snRNP recruitment. The high density of AT nucleotide (lower panel) could favor a sharp difference between exon and intron in terms of nucleotide composition bias, which would favor nucleosome positioning on exons. A- or T-rich sequences located upstream of AT exons, as well as the presence of exonic nucleosomes, could locally (at the exon level) slow down RNAPII, favoring synchronization between transcription and splicing. The high density of AT nucleotides increases the probability of generating decoy signals, such as pseudo BPs or SF1- or U2AF2-binding sites. This constraint could be alleviated by the binding of splicing factors (SF, in green) to these decoy signals, thereby enhancing U2 snRNP recruitment

Back to article page