Quality of solution of massively parallel MSA based on metaheuristic social behavioral model

Due to their importance for computational biology, the methods and algorithms for multiple alignment of biological sequences have been the subject of continuous intensive research and innovation in the last 30 years. Within a PRACE research project we have designed and implemented metaheuristic algorithm for massively parallel MSA and based on it parallel software tool MSA_BG that has been ported on the European supercomputer Juqueen. In this paper we have investigated the Quality of Solution (QoS) of the parallel algorithm MSA_BG for multiple alignment of biological sequences on GPU-accelerated computing infrastructure based on 3 synthesized benchmarks of genetic sequences comprising the 8 segments of the swine virus AH1N1, the 9 genes of the severe acute respiratory syndrome coronavirus 2 isolate SARS-CoV-2/human as well as the human genes BRCA1 and BRCA2 associated with the breast cancer issue. The analysis of the experimental results show that the QoS of the individual runs fall in the range of+10%, except for the case study of BRCA1 and BRCA2 where the range is within ∓1%.