PrionScan, an online database of predicted prion domains in complete proteomes
PrionScan is a database of prion domains predicted for all publicly available proteomes. Based on amino acid propensities previously observed in prion domains experimentally tested in yeast by Alberti et al., 2009, we have built a probabilistic model, benchmarked to handle large sequence databases, and used it to predict prion-like sequence stretches in all the proteomes available annotated in UniprotKB (Angarica et al., 2013).
This site offers two functionalities:
SEARCH the Database. You can make simple searches to find out whether a specific protein contains putative prion domains (using UniprotKB identifier or accession number) or to retrieve all the predicted prions in a given organism. More complex searches combining information from multiple columns can retrieve, for example, all the putative prions from a given organism that exert a specific function or resides in a given cellular component (as described in Gene Ontology). The predictions included in our database correspond to the following UniprotKB database update:
UniProtKB/Swiss-Prot Release 2014_11 of 26-Nov-2014
UniProtKB/TrEMBL Release 2014_11 of 26-Nov-2014
RUN a prediction. Users can process their own sequences (e.g. synthetic sequences, mutants, yet-to-annotate proteins) to make predictions of prion domains according to our model. The Sequence Analysis functionality allows users to paste sequences in FASTA format or to upload files with a large number of sequences (in FASTA format, flat files or gzip or bzip2 compressed files). The cutoff value for prediction can be modified.