Researchers have identified a previously overlooked vulnerability in the human genome where the first 100 base pairs of genes experience mutations at a 35% higher rate than expected. These changes, often occurring in early embryonic development, can lead to mosaic mutations that are passed to offspring and contribute to diseases like cancer. The discovery, published in Nature Communications, calls for updates to genetic models to better detect such variants.
A team led by Dr. Donate Weghorn from the Centre for Genomic Regulation in Barcelona has revealed that transcription start sites—the points where cells begin copying DNA into RNA—are particularly susceptible to genetic alterations. According to their study, released on November 26 in Nature Communications, these initial 100 base pairs accumulate mutations 35% more frequently than chance would predict. "These sequences are extremely prone to mutations and rank among the most functionally important regions in the entire human genome, together with protein-coding sequences," Dr. Weghorn explained.
The mutations primarily emerge during the rapid cell divisions shortly after conception, resulting in mosaic patterns where only some cells carry the change. Such mosaics can evade detection because they do not affect the entire body, yet parents without symptoms may transmit them via eggs or sperm. Offspring inheriting these would have them in every cell, potentially causing health issues.
To reach this conclusion, the researchers analyzed transcription start sites in 150,000 genomes from the UK Biobank, 75,000 from the Genome Aggregation Database (gnomAD), and data from eleven family studies on mosaics. They found excess mutations concentrated in genes linked to cancer, brain function, and abnormal limb development. Rare, recent variants showed the strongest signals, diminishing in common ones, suggesting natural selection eliminates harmful changes over generations.
This hotspot challenges current mutational models, which underestimate baseline mutation rates in these areas. Dr. Weghorn noted that without adjustments, models might misinterpret data, such as expecting 10 mutations but observing 50, when the true baseline is 80—indicating selection at work. Traditional studies focusing on child-specific mutations miss mosaics, creating blind spots. Solutions include examining mutation co-occurrence or revisiting discarded variants near affected start sites.
The vulnerability stems from the chaotic nature of transcription initiation, where machinery pauses, reverses, or exposes DNA during early development. While repair mechanisms exist, the rush of cell growth allows some errors to persist, adding a new germline mutation source alongside known causes like replication errors.