Back to catalogue >>

Reference: Bromberg Y., Tachdav G., Rost B. SNAP predicts effect of mutations on protein function. Bioinformatics (2008) 24, 2397-2398.
Hosted: (http://www.rostlab.org/services/SNAP/) Developed by the Rost lab.

Summary: SNAP uses neural networks (NNs) to predict the severity of missense variants. It segregates ‘neutral’ and ‘non-neutral’ changes based on protein structural features.

Methodology:
• Structural features such as secondary structure, solvent accessibility and conservation are used to characterise a position in a protein.
• Once the NN has made a prediction, a reliability index is provided that loosely correlates with the severity of the variant. Input:
• The query sequence must be inserted by the user.
• Minimum reliability index: SNAP will report only predictions with reliability over a set threshold. Default=0, all are reported.
• Minimum expected accuracy: SNAP will only report predictions with accuracy over a set threshold. Default = 50%

Output:
• A prediction of ‘neutral’ or ‘non-neutral’ is given, along with a reliability index and expected accuracy.