Fig. 7

Computational performance for different tasks. Biotite and its extension packages are compared to other software in terms of computation time for selected tasks. The respective software is given on top of each bar. Each task was run 100 times and the average was taken, if not specified otherwise. k-mer index: KmerTable instantiation vs. mmseqs createindex for Swiss-Prot dataset. Repeat masking was omitted. Computations were performed using a single thread. Due to the high run time, this task was run only once. Alignment search: The workflow from ‘Identification of homologous sequences’ vs. mmseqs easy-search. k-mer indexing was not included in the time measurement. Computations were performed using a single thread. Instead of running mmseqs easy-search multiple times, it was run once with the according number of query sequence copies, to get a more realistic application scenario. MSA: align_multiple() versus clustalw -align [69] for 200 sequences from SCOP [70] globin family. Calculation of pairwise sequence distances and the guide tree is included. The task was run ten times. Hydrogen prediction: Hydride add_hydrogen() and relax_hydrogen() vs. gmx pdb2gmx [71] for a hemoglobin tetramer. Hydrogen bonds: hbond() vs. gmx hbond for a hemoglobin tetramer. ANM: Hessian calculation of an ANM in Springcraft vs. ProDy [72] for a hemoglobin tetramer