Back to catalogue >>

Reference: Calabrese R., Capriotti E., Fariselli P., Martelli P.L., Casadio R. Functional annotations improve the predictive score of human disease-related mutations in proteins. Hum Mut.(2009) 30 1237-1244.
Hosted: Developed in the Biocomputing group at the University of Bologna, it is hosted by a server of the Structural Bioinformatics Unit at the University of the Balearic Islands (UIB). (http://snps.uib.es/snps-and-go/)

Summary:
SNPs&GO uses a support vector machine (SVM) to determine whether variants are disease-associated or not. It is similar in methodology to PhD-SNP, having been developed by the same people.

Methodology:
Parameters integrated into the SVM include:
• The sequence change of the substitution in association with local sequence environment.
• The sequence profile at the site in an alignment with BLAST hits.
• Four measures of output prediction from PANTHER.
• Functional gene ontology (GO) from the query gene and their parents.

Input:
• The user must provide the protein sequence or Swiss-Prot ID.
• Optionally, any known GO terms associated with the gene can be entered.
• By default, the SVM uses all parameters listed above to make the prediction. The option is also available to get predictions from each parameter separately (i.e. predictions using sequence profile, GO and PANTHER individually as well as the combined SVM prediction for SNPs&GO).

Output:
Each substitution is classified as ‘disease’ related or ‘neutral’ and provided with a probability (neutral if <0.5) and reliability index.