New Algorithm Identifies Disease-Linked Variants in Non-Coding Human Genome Regions

Researchers from the Children's Hospital of Philadelphia (CHOP) and the Perelman School of Medicine at the University of Pennsylvania have developed an innovative algorithm to detect genetic variants in non-coding regions of the human genome that may contribute to disease risk. This breakthrough approach focuses on the vast portions of DNA that do not code for proteins but play crucial roles in regulating gene expression.

While the human genome consists of over 98% non-coding sequences, identifying disease-associated variants within these areas has been historically challenging. Traditional genome-wide association studies (GWAS) have pinpointed broad regions linked to conditions, but isolating the exact variants responsible often remains difficult. Many of these variants are located near transcription factor binding motifs—specific DNA regions where proteins involved in gene regulation, called transcription factors, attach to control gene activity.

The research team employed ATAC-seq, a technique that maps accessible, "open" regions of the genome, and combined it with a deep learning method called PRINT, capable of identifying footprints left by DNA-bound proteins. Analyzing data from 170 human liver samples, they identified 809 specific locations, known as footprint quantitative trait loci (footprint QTLs), which associate genetic variants with the strength of transcription factor binding.

This method allows scientists to see how different genetic variants affect the binding of transcription factors at certain sites, providing insights into how genetic variations influence gene regulation and potentially lead to disease. Dr. Struan F.A. Grant explained that this approach is akin to distinguishing the real culprit in a lineup of suspects, by pinpointing the precise DNA footprint of the disease-causing variant.

The researchers aim to extend this technique to other tissues and organ samples to identify variants that drive various common diseases. According to first author Max Dudek, this approach offers a new way to uncover causative noncoding variants, which could eventually lead to novel treatments.

Published in the American Journal of Human Genetics, this research marks a significant step towards understanding the complex regulatory code of the human genome and its impact on health and disease.

Source: https://medicalxpress.com/news/2025-04-algorithm-potential-disease-variants-coding.html