Correlated evolution of positions within mammalian cis elements

Area of Science:

• Genomics
• Bioinformatics
• Molecular Biology

Background:

• Transcription factor (TF) binding site models, like positional weight matrices (PWMs), typically assume independence between nucleotide positions.
• Evidence suggests that this independence assumption may not always hold true for TF-DNA interactions.
• Understanding interposition dependence is crucial for accurate modeling of transcriptional regulation.

Purpose of the Study:

• To develop and apply a maximum-likelihood (ML) approach to infer correlated evolution between positions within TF binding sites.
• To investigate the prevalence and characteristics of interposition dependence in cis-regulatory elements across mammalian genomes.
• To explore the implications of correlated evolution for refining TF-DNA binding models.

Main Methods:

• Developed a maximum-likelihood (ML) method to detect correlated evolution between any two positions in a TF binding site motif.
• Applied the ML method to a multiple alignment of 5 mammalian genomes, analyzing genome-wide putative cis elements in human promoters.
• Examined the relationship between interposition distance and dependence, evolutionary constraint, and similarity across related TFs.

Main Results:

• A prevalence of correlated evolution was detected within cis elements across mammalian genomes.
• Interdependence between positions decreased with increasing physical distance.
• Interdependent positions were found to be evolutionarily more constrained, with conserved patterns across related transcription factors.

Conclusions:

• Nucleotide positions within mammalian TF binding sites exhibit significant correlated evolution, indicating interposition dependence.
• This interdependence likely arises from context-dependent nucleotide preferences within cis elements, rather than solely from mutational biases.
• The developed methodology can refine TF-DNA binding models and is applicable to other non-coding functional elements.