On average, the genomes of two humans vary by 3 million nucleotide variants.
Although most of these variations are not life-threatening, some may cause serious health disorders. There are more than 7000 rare genetic diseases, most of which cannot be treated. However, it is possible to test the genomes during preconception, prenatal, and early childhood periods. Although the early diagnosis of such diseases is crucial, knowing the precise effect of most novel variants is a challenge. It is infeasible to perform billions of unique experiments in laboratories to cover the entire theoretical nucleotide variants.
To cope with that, the project developed an algorithm, PHACT, that computationally predicts the effect of nucleotide variants in the protein-coding regions. PHACT outperforms its competitors by benefiting phylogenetic trees that are computationally expensive to generate. The project has shown capabilities for efficient parallelization in MeluXina through a benchmark study in the scope of EuroHPC. Here, they requested more computational resources to cover the entire set of all possible amino acid substitutions in all human proteins.
The anticipation is that the results will be effectively used not only by other research groups but also by clinicians interested in predicting the effect of such mutations.
Sabanci University, Turkey.