Predict Sites

This program searches for matches to a plant-specific target sequence for N-terminal myristoylation. The search uses a profile hidden Markov model, 19 residues long, that has been trained on known and probably N-myristoylated plant protein sequences. Details of the model development are described briefly on the validation data page, and in more detail in the published paper (Podell and Gribskov, 2004).

For each query sequence, the program reports the highest scoring match, plus any additional matches with scores greater than the threshold cutoff value (0.55). A P-value (log probability) is calculated for each match based on score frequencies on all predicted proteins in the Arabidopsis thaliana genome.

Each query should contain one or more sequences in FASTA format. Minimum sequence length required for valid results is at least 20 residues. This form is limited to 1000 sequences per query - please contact us if you would like to run larger data sets.