The DeCypherHMM™ algorithm is our version of the popular HMMER 2.0 algorithm, which utilizes Hidden Markov Models to compare protein sequences. This sensitive algorithm allows one to compare distantly related sequences and in the process, identify the domain structure of protein families. The DeCypherHMM™ suite is available as a DeCypher module in our DeCypher® systems.
- Model-based analysis helps characterize multi-domain proteins.
- Includes both hmmsearch and hmmpfam
- Frameshift-tolerance helps you classify evolutionarily divergent proteins.
- Ideal for classifying novel proteins for metagenomics studies.
- Easily build and expand model databases from your sequence data.
- Can return multiple hits per protein domain using J-state processing.
Coming in 2014 - Accelerated HMMER3.0
When HMMER3.0 was introduced as a replacement to HMMER2.0, several significant algorithmic changes were introduced. Perhaps most importantly, HMMER3.0 is a heuristic algorithm, whereas its predecessor, HMMER2.0, was a dynamic programming algorithm. These changes significantly reduce the amount of data that HMMER3.0 processes and accordingly, performance improves dramatically.
TimeLogic has chosen to emphasize heuristic algorithm performance with our latest FPGA-based DeCypher J-Series Similarity Search Engine Accelerator and as a result, we are well positioned to introduce the only accelerated implementation of the HMMER3.0 algorithm. DeCypherHMM3™ will be available in 2014.