Application of AI in Precision medicine

AI in precision medicine

Precision medicine is a systemic healthcare process that cumulatively applies multiple technologies to individually diagnose and treat patients based on their own biological make up. The dawn of the millennium marked the completion of the human genome project, which brought about a surge of development in the field of genomics and its application in diverse fields. These large-scale developments in ‘omics’ have resulted in a rapid evolution of precision medicine which shows significant promise in improving population health. However, the field is still in its infancy and has its own challenges. 

The basis for personalizing health and lifestyle recommendations based on genomic information is one aspect of precision medicine that is quickly gaining ground due to the off-shoot of many direct-to-consumer (DTC) companies. This preventative healthcare space encompassing the field of nutrigenomics is believed to offer health suggestions and knowledge of risk traits, based on natural variations occurring in the genome called single nucleotide polymorphisms (SNPs). Lifestyle modifications in accordance with the suggestions may have the potential to eventually reduce the burden of risk. However, due to the void in the number of studies, translating the findings effectively using a single SNP approach is viewed with critical lenses. Understandably so, as most biological functions involve the interaction of multiple enzymes and therefore many genes. In addition to genetic factors, epigenetics and the microbiome also contribute to the final phenotype, further confounding data interpretation.

That said, the major point of concern for precision medicine is the lack of efficient ways to translate large research data into meaningful application for the population at large. Although DNA sequencing has been at the forefront of genomics, currently most of the genomics based technology has the ability to only short read genomic sequences. In addition, there are no solutions to analyze the unknown risk contribution of repetitive regions and structural variations. Advances made in a similarly exciting field, Artificial intelligence (AI), offers promise. AI works on the ability to train high performance computers with artificial learning algorithms to identify patterns and variations in multidimensional datasets, that are usually impossible for a human to decipher. The core concept is to employ learning strategies to identify similar, and also unique patterns in other individuals to predict and optimize the interpretation. Observations indicate that the most promising approach for analysis and identification of polygenic risk scores relies on employing neural network driven machine learning algorithms rather than using polynomial algorithms. 

Recently machine learning has also been adopted to read long stretches of DNA fragments from digital electronic signalling data. Long read technologies will be able to resolve the complexity of repetitive regions in the genome and detect complex structural variants. The nano-pore sequencing technology in particular, has begun to use a neural network-based deep learning method base calling the DNA sequence, which was seen to have an accuracy of over 98% and can produce mega base long DNA reads.

The other critical challenges associated with genomics include making functional sense out of large data and understanding the inherent complexity in classifying mutations according to their clinical relevance, due to the largely unknown penetrance of SNPs. Moreover, low penetrance SNPs are much more common than ones with higher penetrance and most of these variants are non-coding in our genomes. Therefore, determining pathogenicity of rare or common non-coding variants still requires major advances in genomics. Furthermore, many penetrant variants are known to have more than one clinical manifestation, known as pleiotropy, and many diagnoses are characterized by variable presentation

In line with these requirements, attempts have been undertaken to use AI in the clinical classification of genomic variation. The ability of AI to analyze multidimensional biological data and the use of several approaches will potentially help to decipher the pathogenicity. The basis for these learning algorithms include, analyzing other contributory factors like characterization of non-coding variants, splicing code, DNA/RNA binding proteins and non-coding RNA (ncRNA) using large scale molecular datasets. 

Preventive healthcare is increasingly being favoured by healthcare providers and consumers alike and has the potential to revolutionize the perception of health. As research integrating AI into precision medicine progresses in leaps and bounds, it is evident that application of AI in precision medicine holds significant promise for the future.

Editor: Sindhu Menon
Illustrator: Annmary Erinjeri

Visit our “News & Views” section for latest developments in AI-powered digital healthcare.

Leave a Reply