Gene expression models based on transcription factor binding events confer insight into functional cis-regulatory...

Area of Science:

• Genomics
• Computational Biology
• Molecular Biology

Background:

• Deciphering cis-regulatory variants' functional roles is a key challenge in genome analysis.
• Altered transcription factor (TF) binding is a hypothesized mechanism impacting gene expression.
• A computational framework is needed to quantify these mechanistic contributions.

Purpose of the Study:

• To present TF2Exp, a gene-based computational framework.
• To predict the impact of altered TF-binding events on gene expression.
• To provide a tool for interpreting functional roles of cis-regulatory variants.

Main Methods:

• Developed a gene-based framework (TF2Exp) for predicting gene expression.
• Utilized data from lymphoblastoid cell lines for model training and validation.
• Incorporated features such as DNase I hypersensitive, CTCF-bound, and TF-bound regions.

Main Results:

• TF2Exp successfully predicted expression levels for 3196 genes.
• Model performance was comparable to common variant-based models in cross-validation and external validation.
• Combining TF alteration and common variant features improved predictive performance.
• TF2Exp can evaluate variants in linkage disequilibrium and uncommon variants, increasing predictable genes.

Conclusions:

• TF2Exp offers a novel framework for understanding cis-regulatory variant functions.
• The framework enhances the interpretation of gene expression changes driven by TF-binding alterations.
• TF2Exp is a significant step towards functional interpretation of the human genome.