Predicting Protein Ligand Binding Sites with Structure Alignment Method on Hadoop

The application of Hadoop in Structural Bioinformatics

“…In molecular docking, [16] observed a fall in parallel efficiency of 55.3% (99.1% -43.8%) when scaling from 2 nodes to 32 nodes. In structural alignment, [39] observed that performance degraded slightly after 8 mappers was increased to 30. Furthermore, this trend has also been observed on the MS Azure platform we have discussed for comparison -in scaling Cloud4Psi, also a structural alignment application [51], observed that horizontal scaling resulted in performance degradation as a result of an increase in nodes sharing a virtual disk, and that vertical scaling resulted in performance degradation as a result of increased CPU utilisation (due to more processes running per node).…”

“…A variety of of algorithms have been developed to perform structural alignment of proteins, such as, for example, STRUCTAL (Structural Analysis Algorithm), DALI (Distant Alignment) [22], CE (Combinatorial Extension) [69], VAST (Vector Alignment Search Tool) [19], and FATCAT (Flexible structure Alignment by Chaining Aligned fragment pairs allowing Twists) [82], SSAP (Sequential Structure Alignment Program) [57], and MUSTANG (Multiple Sequence Alignment Algorithm) [32]. The technique has been applied to the study of protein binding sites, and solvent exposed surfaces (these effect the energetics of protein-ligand conformations) [42,33,39].…”

“…There is also a need to apply such techniques at scale. For example, Hadoop has been employed by [39] who implemented structural alignment for binding site prediction. Using a test sets of 200 and 48 ligand-protein complexes, they were able to achieve 93% and 98% accuracy, respectively, and were able to improve the efficiency of the experiment by exploiting parallelisation.…”

The application of Hadoop in structural bioinformatics

“…A variety of of algorithms have been developed to perform structural alignment of proteins, such [77], CE (Combinatorial Extension) [78], VAST (Vector Alignment Search Tool) [75], and FATCAT (Flexible structure Alignment by Chaining Aligned fragment pairs allowing Twists) [79], SSAP (Sequential Structure Alignment Program) [80], and MUSTANG (Multiple Sequence Alignment Algorithm) [81]. The technique has been applied to the study of protein binding sites, and solvent exposed surfaces (these effect the energetics of protein-ligand conformations) [82,83,84].…”

“…There is also a need to apply such techniques at scale. For example, Hadoop has been employed by [84] who implemented structural alignment for binding site prediction. Using test sets of 200 and 48 ligand-protein complexes, they were able to achieve 93% and 98% accuracy, respectively, and were able to improve the efficiency of the experiment by exploiting parallelisation.…”