Developing an AI solution that provides predictive tests for HIV
A world leading infectious disease group engaged us to develop an AI solution that provides predictive responsiveness tests for HIV natural immunity on individuals genotypes - specifically for the sex worker population in Kenya.
Use single-nucleotide polymorphism (SNP) genome data to develop a model which will permit the creation of multi-snip algorithms to explain the variability in HIV immunity and explain HIV natural immunity.
Our gene-path application has revealed opportunities for public and private organizations in fields such as metabolic pathway analysis, bio-sample analysis and bio-security. Working with the world renowned infectious disease and bio-terrorism experts, we have developed AI tools to identify SNPs which underlie differences in people’s susceptibility to diseases and support disease / bio-terrorism agent classification and identification.
Our AI solution includes a causal Machine-learning model using Bayesian Networks (probabilistic graphical modelling), utilizing digitized genetic and epigenetic data and disease incidence. This AI tools permit the creation of a multi-SNP models to explain the variability in disease surveillance, detection and impacts of external and environment factors on health outcomes. The solution is integrated with image based analysis, such as Digital Pathology (DP) and Virtual microscopy, to extend diagnostic capabilities to tissue or image classification: normal, benign or malignant. Genetic, epigenetic, and phenotype information can be linked to KEGG, BRITE, DAVID, and other bioinformatic tools, enabling government researchers to quickly leverage their observations into models and connect them to domain relevant pathway analysis.
We developed a proprietary PGM model for genetic analysis, including SNP and RNA expression data. The solution narrowed down the ‘pattern’ of genes that are involved with HIV natural immunity. Meanwhile, proprietary algorithms developed and other AI models are being deployed with the team that are undertaking a vaccine study in monkeys; predicting which factors are most predictive of survival during trials.
Other Use Cases