The service provides statistical predictions of membrane protein structures based on a hidden Markov model. It accepts amino acid sequences as input and generates expected transmembrane helices implied by the input. The same functionality is available as installed software, or through an interactive web interface or SOAP API.
API methods support submission of up to 2,000 amino acid sequences in standard alphabetic notation. Methods adjust for unexpected or missing sequence values in applying statistical analysis to generate predictions of expected protein structures. Methods also support monitoring the processing queue and retrieving results when the job completes.