Hidden Markov Models Offer a Powerful Approach for Understanding Gene Regulation Mechanisms Relevant for Organ Transplantation

The human genome contains sequences of DNA enriched in cytosine-guanine dinucleotides known as CpG islands (CGIs). CGIs play a crucial role in gene regulation and expression, making them an important target for genetic therapies. In this article, hidden Markov models (HMMs) and adaptive window techniques (AWTs) were used to identify CGIs in MUC5B and DSP genes. Both genes are associated with idiopathic pulmonary fibrosis, a progressive pulmonary disease that leads to a lung transplant. The University of California, Santa Cruz Genome Browser was used to obtain the MUC5B and DSP gene sequences and predefined CGI locations. The HMM and AWT algorithms were developed using Python version 3.11.5, and the outcomes analyzed were sensitivity, specificity, computational memory, and runtime. Both HMM and AWT exhibited high specificity; however, HMM was more accurate than AWT for both genes, 99% versus 96%, respectively. The HMM sensitivity was higher for both MUC5B and DSP genes (87% and 88%) compared with only 58% for MUC5B and 57% for DSP with AWT. Regarding computational efficiency, AWT was faster and required less memory than HMM for both genes. By accurately detecting CpG-rich regions, HMM offers a powerful approach to understanding gene regulation mechanisms. This could pave the way for more precise therapeutic interventions, enabling targeted treatment strategies for a range of genetic disorders, including idiopathic pulmonary fibrosis, improving patient outcomes, and advancing personalized medicine.