A common misconception is that single gene variations exist for specific traits. In fact, the majority of traits and non-communicable diseases are complex and multigenic: caused by a combined effect of many genes, transcription and external factors.

Coronary artery disease, Type 2 diabetes, Alzheimer's disease, and cancer are all examples of complex disease and are among the leading causes of mortality in Australia and most developed countries. Their genetic complexity makes them more difficult to predict and manage compared to single-gene disorders. This presents a significant opportunity for computational tools to analyse the vast genomic data and help manage polygenic disease.

Our new article introduces a novel method for predicting an individual's risk of hypercholesterolemia and hypertension from a whole genome sequence. This approach involves two deep neural networks working in series to analyse both indel variants and single nucleotide variants (SNVs) to predict risk. Unlike conventional methods that focus solely on SNVs, our approach incorporates indel mutations, which are short insertions or deletions of DNA. The new method has improved prediction accuracy by 21% as compared to previous methods. 

While initially tested on cardiovascular disease, this algorithm shows great potential to predict other polygenic diseases, such as Type 2 diabetes, Alzheimer's, and cancer. At 23Strands, we aim to leverage advances in genomics to provide personalised insights into patients' health. This new algorithm will expand our ability to predict complex polygenic diseases and empower identified high-risk patients to make more informed decisions about their future health. 
 
Soo, Z., Lin, H., Yang, Y., Grosser, M., Zhang, Y. and Lu, J., 2023, November. Deep Neural Network-Empowered Polygenic Disease Prediction on Cardiovascular Diseases. In 2023 18th International Conference on Intelligent Systems and Knowledge Engineering (ISKE) (pp. 309-315). IEEE. https://doi.org/10.1109/ISKE60036.2023.10481444