SUPERFAMILY Model Library

The model library is created by building models using 6 iterations of the T99 procedure for every member of the PDB (filtered to 95% sequence identity). These models are then scored against the whole of the PDB to detect models with problems, and to assess the score thresholds for the genome assignments.
Some degree of human intervention and effort goes into improving these models to try to make this library the best possible.

Online sequence search

To search sequences against the HMM library go to the sequence search page and enter sequences for searching. Sequences can be entered in 2 ways:

• There is a text box in which you can paste sequences in fasta format. You may also paste or type a single sequence in no particular format.
• You may upload a file from your computer which contains sequences in fasta format.

Download model library

The models which make up the library are available to download in SAM and HMMER format (converted using this script).

Please register for a license to download the SUPERFAMILY models. Licenses are free for academic and commercial purposes.

Read the documentation for more information and help using the hidden Markov models and associated scripts.