Hidden markov model
A hidden markov model (HMM) is often used in bioinformatics to model a biological pattern in sequences. It can hereby take into account for each position of a pattern the possibility of which residues is at that position of the sequence, the possibility of a deletion and the possibility of an insertion. The three possibilities are called states. A HMM also includes the chance for going to one state to another ('transition probabilities). Hence, biological patterns can be modeled in a rather complex way using HMMs, leading to sensitive searches.
A hidden markov model is a machine learning technique, meaning that the model is automatically generated by inputting sequences (e.g. in the form of a multiple sequence alignment).
Hence, a HMM can be used in the case:
- you want to detect in a sequence a domain/motif
- you want to search a sequence database with a multiple sequence alignment
- HMMER3 - A very user-friendly and powerful interface to use HMM!