A tool to predict the effect of missense variants
Polyphen, now available in its version Polyphen-2, predicts the impact of a missense variant (also referred to as nsSNP: non synonymous single nucleotide polymorphism) based on (1) protein sequence (2) phylogenetic information and (3) structural information. So Polyphen is able to do the so called functional annotation of missense variants. The software actually looks if the mutation is falling within a protein domain essential for the binding to other molecules of for the formation of the secondary/tertiary structure. In particularly Polyphen-2 looks at putative disulfide bonds, active sites, binding sites and transmembrane domains and makes computations on 3D models of the protein structure. Polyphen-2 also looks at homologous proteins to see if the identified missense mutation has been observed in other proteins of the same family.
How to use it
Accessing Polyphen-2 predictions is possible through the Polyphen-2 website. Many bioinformatic software (licensed and open-source) also redirect to this page. An input mask is available on the homepage. The query can be initiated by giving the variant coordinates (for instance the rs number). The results won’t be immediately available, as your query will be queued and processed as soon as possible. By clicking on the refresh button one can see the status of the analysis (an example in the image below).
The analysis can also take several minutes, depending on how busy the Polyphen-2 server is. In the example shown the analysis took about 45 minutes (see image), but waiting times are usually shorter. When the job’s ready, we get a variant score represented as bar colored from green (score 0.0 – predicted as benign/neutral) to red (score 1.0 – predicted as pathogenic).
It is important to highlight that Polyphen-2 can also analyze a large number of missense variants at once (batch analysis) through its web server. There’s a dedicated page for this on the Polyphen-2 website. Detailed explanations on how to use the single variant or batch analysis functions are available in the Polyphen-2 flagship publication (PMID: 23315928, free).
Breda Genetics and Polyphen-2
Polyphen-2 is so far considered one of the most powerful and reliable tools to predict the effect of a missense variant. Thanks to specific algorithms embedded in our bioinformatics pipelines, any missense variant detected in Breda Genetics‘ exome and genome sequencing is automatically analyzed with Polyphen-2. Breda Genetics utilizes also additional in silico prediction tools for missense variants, such as SIFT and Mutation Taster. In patient’s interest, our interpretation of previously unreported missense variants is always very conservative. So no in silico prediction alone is for us enough to definitively assess the meaning of a variant. Other evidences, primarily family testing, are necessary to confidently assess the meaning of such variants.