Which tool performs better for domain detections? What is your experience with these two?
Just a few of additional considerations: Interpro database contains up to 11 different specific databases of protein motifs,signatures, sites, which include PFAM database, so when you run InterproScan you might obtain more annotations than running a single PFAM search, however if annotations look generally more abundant they may look also redundant, which is not bad at all as redundant, or more correctly, convergent annotations using different approaches or databases elegantly displayed in a single graphic when using Interproscan web service may give you either a more comprehensive or at least confirmatory functional annotations of your gene or protein. What I like with Interproscan is exactly this, the integration of different annotations from different kind of protein signatures databases and that can be displayed together with links to all specific databases and in addition to GO terms.
On the other hand PFAM has a subset of highly curated entries (PFAM-A) which makes it interesting for more confident analysis of protein domains and some query tool parameters can also be adjusted at their web-based interface. Interestingly it has a complementary interaction tool (iPFAM) based on real interaction evidences from structural data between different protein domains, as well as links to structural information of the domains (PDB). PFAM queries can be done using HMMER tools as it has been built with the same algorithms: I agree with Gustavo that Hmmer search tools are very fast and efficient, even for HTS derived datasets, and much more robust than typical blastx-based annotations…
Interproscan can also be downloaded if you have a large dataset of sequences as queries (transcriptomic / genomic data).
If you want to know more about each database and related tools take a look to these two 30 min-long EBI's online tutorials:
http://www.ebi.ac.uk/training/online/course/interpro-quick-tour
http://www.ebi.ac.uk/training/online/course/pfam-quick-tour
Hope other people with the same inquiry will find it helpful
Best
http://www.ebi.ac.uk/training/online/course/interpro-quick-tour
http://www.ebi.ac.uk/training/online/course/pfam-quick-tour
Both Interpro and PFAM are databases, not tools for domain detection. I suppose you are talking about InterproScan and HMMer, that are the commonly used tools to scan a protein against those databases, respectively,
InterproScan will search your query against diverse databases of protein domains (including PFAM, using HMMer), motifs, signatures and families. It will also predict transmembrane regions and signal peptides. Shortly, Interproscan outputs what a search against PFAM would output and much more information.
The disadvantage is the runtime. Interproscan webservice can be very slow if you need to analyse thousands of proteins, but you can download and install the whole suite locally (need some time to configure it to run in a local computer cluster or in the cloud).
If time is a limitant, HMMER v3 wwill fastly scan thousands of proteins for PFAM domains in few hours.
In my experience, for the same protein Pfam can miss the domain that can be seen in Interpro results as Pfam-domains.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics
- Bioinformatic Tools