An article describing MILANO is now published in BMC Bioinformatics: http://www.biomedcentral.com/1471-2105/6/12/
MILANO (http://milano.md.huji.ac.il) is a web based tool for automatic literature searches on lists of genes. It helps in identifying significant genes out of a list of, e.g. upregulated genes from a microarray experiment, by cross-searching the genes with user-provided terms. MILANO accepts LocusLink (now Entrez Gene) Id's and performs synonym expansion on them and searches literature databases, PubMed and GeneRIF. It is similar to previously published search tools (such as pubmatrix, geneinfo, microgenie) but has several advantages - mainly the ability to search the GeneRIF database, which is a collection of summarized gene-specific articles. GeneRIF searches are performed online and are relatively quick. PubMed searches are queued to run according to NCBI terms of use (slowly...)
MILANO is written in perl (CGI), runs on linux and uses gnqs for queueing.
I'm planning to make it open-source once it's published.
Comments will be gladly accepted!
Ran.
Comments
Admin
I haven't done a word for word comparison, but I assume this story is a duplicate of the story in the queue ? I will delete the one in the queue if that is okay, if not let me know.
Milano looks interesting, but unfortunately I don't have much expertise in this area to comment seriously. Do you plan on developing it further ?
Further development
Currently Milano is in 'alpha'. I plan to release the source (with a proper open-source license) once I clean it up .
Further advancements will include accepting different types of accession numbers, such as Affymetrix probeset IDs. Currently I use Locuslink/Gene Id's because I feel they have the most meaning for this type of search, and because GeneRIF is based on these genes. Also, most microarray vendors supply LocusLink IDs in their annotations.
Feature suggestions will be gladly accepted!
duplicate
This story appeared recently then disappeared - I promoted it back to the front page. Don't know if that caused any problems.
I have little/no microarray knowledge, but a problem I have with this kind of approach is that it's not clear what the service is meant to do or what I should be typing in as search terms. Perhaps I would have a better idea if I did microarray experiments. I assume the idea is to correlate gene annotation data with expression data? In which case I can think of better ways than small, limited searches - but then I am a proponent of "all against all" approaches.
Application for Milano
Let me give you an example. When performing a microarray experiment in which you test expression response to p53 (a DNA damage-related transcription factor and Very Important Protein), you get a list of 150 genes that were significantly up- or down-regulated. Now you want to find information about these genes. Milano enables you to perform automatic searches in the literature for these genes and cross them with experiment-specific search terms.
For example, searching these genes with the term 'p53', will give you positive hits on genes that co-appeared in articles with their putative regulator (as detected in the microarray experiment). This enables you to find known p53 targets, or even regulatory feed-back loops - if you find regulators of p53. Other intersting applications can be thought of.