Basel Abu-Jamous, Director of Computational Biology at Nucleome Therapeutics, unpacked the challenges of mapping non-coding region variants to specific genes. This is a field typically overlooked in genomics, with most companies steering clear of the dark genome.
The principal feats associated with non-coding regions are determining whether a variant affects gene regulation and identifying which gene(s) a non-coding variant regulates, especially when distant from the gene. Furthermore, the enhancers and non-coding regions of the genome are key to the differentiation between different cell types. Abu-Jamous elaborated that the same gene might be regulated by different enhancers in different cell types.
Nucleome Therapeutics uses a proprietary 3D genomics technique called Nucleome Capture-C to map interactions between genome points, achieving single base pair resolution. This technique helps identify which genes are regulated by specific non-coding variants.
Abu-Jamous sought to validate the various methods using ATAC-seq to assess chromatin openness and RNA-seq to measure gene expression changes. These techniques confirm the functional impact of variants identified by machine learning and 3D genomics.
The company employs in-house machine learning models to predict the functionality of non-coding variants, including their impact on chromatin openness and gene expression. These models are cell type-specific, enhancing the accuracy of predictions. Nucleome Therapeutics focuses on autoimmune diseases, but its tech is more disease-agnostic meaning it could be applied across a range of therapeutic areas.
Understanding and uncovering which variants interact with which promoters of genes at a genome-wide level is highly complex due to the amount of ambiguity and it may generate a lot of false positives. One critical advantage of the Nucleome Capture C is its ability to be precise at scale. Even on a large scale one can use the technology to observe where certain points of the genome interact with other points of the genome. The technology can achieve single base pair resolution if required.
Nucleome Capture C can be leveraged in drug target identification and validation. For example, GWAS identified and validated novel therapeutic targets from non-coding GWAS hits. Then, machine learning methods pinpointed SNPs as potentially predicted to be functional and influence chromatin openness. Then 3D Genomics finds the genes that the functional variants interact with. In summary, the methods employed at Nucleome Therapeutics identify novel drug targets by linking disease-associated variants to gene dysregulation, which can then be validated and prioritised for further research.
To wrap up, the integration of machine learning, high-resolution 3D genomics, and allele-specific validation assays provides a powerful framework for decoding the regulatory genome. This approach boosts drug discovery pipelines and makes valuable contributions to scientific understanding of gene regulation in health and disease.
