Here we provide a list of all pre-computed predictions with their logistic regression scores for human, fly, and yeast. Datasets are available below in tab-delimited format. All sequences per species are provided as well:

Predictions Species Size Sequences
s2n-all-1.0.tar.gz all 9.65 MB s2n-all.seq.tar.gz Drosophila melanogaster 6.48 MB s2n-dmela.seq.tar.gz Homo sapiens 16.02 MB s2n-hs.seq.tar.gz Saccharomyces cerevisiae 2.11 MB s2n-scere.seq.tar.gz

Web services

Struct2Net provides programmatic access to pre-computed predictions as well as making new ones. Provided here is a set of web services. Documentation on how to use a web service is provided by each script. For example, in the directory of a given script, type $ python search_s2ndb to bring up usage instructions for search_s2ndb

Web ServiceDescription
search_s2ndb Searches over predicted interactions from Struct2Net that have been pre-computed in human, fly, and yeast. Given a query id/name and organism option, this web service returns a list of binary interactions and a confidence score evaluating each interaction.
(coming soon!)
Homology approximation over Struct2Net's pre-computed predictions. Given a query sequence(s), this web service performs a BLAST search over the set of human/fly/yeast proteins and returns the set of predicted interactions for the best-hit homologs.
(coming soon!)
Runs a full-blown prediction, involving both a threading algorithm and machine learning algorithm. Pairs of sequences are threaded onto all templates of a dimer template library. A logistic regression score is then computed for the pair reflecting their probability of interaction.