I am working on integrating pathway relations into a pathway diagram. I have two text mining solutions that I use to find interactions. I am looking for new suggestions. The idea is that I provide a pathway entity (ie. gene/protein, metabolite) and all relations based on the literature are returned.
You might want to take a look at STITCH. It is a database that allows you to query with a gene/protein or a small molecule (e.g. a metabolite) and retrieve the interaction partners. A very large part of the evidence in STITCH comes from text mining, and if you really want to, you can turn off all the other evidence types to exclusively get the results obtained through text mining.
Edit:
If you are using Whatizit, I would suggest that you take a look at the REST API of Reflect. It is designed for speed, so you do not need to worry about hammering the API. If you send too many requests at the same time, some of them will simply be rejected and you will get an error code back.
For gene/protein interactions, try iHop, which has an API described here.
There is Oscar which knows about chemical structures, and used in ChemicalTagger for detecting biological roles, such as solvent roles in this demo, created in a collaboration with the ChEBI team.
You might want to have a look into BER (Biological Entities and Relationships) developed by Temis which provides relationships (including interactions) between proteins/genes aswell as relationships between these and other biological entities (disorders, cells and tissues, treatments, chemical entities, etc....). This solution can be tuned to find relationships only between those entities you are iterested in and/or for the selection of relationship types which you are interested in, and can integrate Oscar if desired. You will find more about Temis at www.temis.com and can contact yoann at yoann.marsac@temis.com for more information on functionalities and deployment.
Since I have asked this question, I also (re)found the following text mining apps: Chilibot chilibot.net) and Agilents literature plugin for Cytoscape to this question/answer serie agilent.be/labs/research/litsearch.html
Other text mining tools are: SciMiner, EbiMed FASTA+ GoPubMed iHOP
;)
Yeah, iHOP would be one of my top open-source tools. PolySearch is another one (it is a little more buggy, but more flexible in the types of queries that it can handle).
http://www.ihop-net.org/UniPub/iHOP/
http://wishart.biology.ualberta.ca/polysearch/
If you have a license, IPA is also an excellent tool for mining the literature (and probably my favorite)
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
I am using Whatizit: http://www.ebi.ac.uk/webservices/whatizit/info.jsf and Phasar: http://www.phasar.cs.ru.nl/. The first works decently, the second has issues with performance and accessibility.
Could you please tell us which are the two text mining solutions that you already have?
I need to add Chilibot (http://www.chilibot.net/) and Agilents literature plugin for Cytoscape to this question/answer serie (http://www.agilent.be/labs/research/litsearch.html)