HMM Performance

Hidden Markov Model and Profilesearch analyses can provide unique benefits over sequence-based analyses. Because they provide more probabilisitcally correct modeling, these searches be more accurate for evolutionarily distant relationships. The PFAM, TIGRfam and Superfamily HMM databases are excellent tools for idenifying the family to which a protein or domain belongs. Searching these databases is dramatically faster on DeCypher than on CPU and cluster-based solutions.

1000 GOS Proteins compared to PFAM 22.0

A CodeQuest biocomputing workstation reduced the runtime of a large HMMpfam Search from 23.67 hours down to 2.5 minutes, for a performance equivalent to 568 CPU cores. This search was performed on a CodeQuest X9 with 2 DeCypher Engines, compared to running HMMer software on a single core 2.33 GHz CPU.

Establishing novel protein families is critical for metagenomics studies. Currently, several institutions employ TimeLogic solutions for their metagenomics projects, including The Scripps Genome Center, DoE Joint Genome Institute, The Salk Institute Max Planck and Argonne National Labs.

Interested in further discussion of our benchmark data? Would you like to see custom benchmarks for your toughest bioinformatics bottlenecks? Contact a TimeLogic account executive today to learn more about CodeQues or DeCypher!