[BioC] Looking for a function

Tim Triche, Jr. tim.triche at gmail.com
Thu Apr 17 18:05:38 CEST 2014


I think Compendia/Life/Thermo still does something like this with Oncomine, but it seems to be a "value add" with the "concepts edition". On the other hand, they do pay professionals to curate the data, which costs money, so...

--t

> On Apr 17, 2014, at 7:07 AM, Kevin Coombes <kevin.r.coombes at gmail.com> wrote:
> 
> A long time ago, a company called InPharmix sold a tool called PDQ_Med that did exactly this. I don't think it ever caught on, but their web site from 2004 is still google-discoverable.
> 
> They actually started from a list of genes (as in the results of a differential expression analysis) and disease terms, and built a connections network on the fly.  A simple form of the algorithm would be
> 
> [1] Search PubMed using a MeSH term (like those in your example) and keep a record of the article-ids that comes back.
> [2] Search PubMed using a gene symbol, and keep a record of the article-ids that come back.
> [3] Intersect the two lists, and analyze the abstracts from the overlap. (InPharmix did some clever text mining stuff that included parsing out abbreviations introduced in the abstract and looking for co-occurrences of gene or disease terms in close proximity, such as in the same or adjacent sentences. They also built networks connecting multiple genes and disease terms.)
> 
> Anyone really interested could try contacting Jim Sluka to see if he's willing (or able) at this point to make any of the code available.
> 
>  -- Kevin
> 
>> On 4/16/2014 5:24 PM, Steve Lianoglou wrote:
>> Hi Jerry,
>> 
>>> On Wed, Apr 16, 2014 at 2:07 PM, Jerry Cholo <jerrycholo at gmail.com> wrote:
>>> Hello,
>>> 
>>> 
>>> 
>>> In Bioconductor, I am looking for a function, in that the input of the
>>> function would be a word such as "inflammation", "fibrosis", or
>>> "apoptosis", and the output of the function should be the list of
>>> significant  "gene symbols" associated with the input words.  The degree of
>>> association will be evaluated by FDR or p-value.  Basically, the function
>>> will use publicly available datasets.  I appreciate if someone provide me
>>> the name of this function.
>> Dollars to donuts: there is no such function.
>> 
>> You could, however, ask what genes are annotated with a certain
>> function (ie. GO term).
>> 
>> For instance, query AmiGO to see what GO terms are associated with inflammation:
>> 
>> http://amigo.geneontology.org/amigo/search/ontology?q=inflammation
>> 
>> One such term is "GO:0002544" (chronic inflammatory response).
>> 
>> Given the GO:XXXX id,  you can ask what genes are associated with it.
>> To start figuring out how to do that, this tutorial will likely help:
>> 
>> http://bioconductor.org/help/workflows/annotation/annotation/
>> 
>> And so, too, would googling for something like "bioconductor query go term"
>> 
>> HTH,
>> -steve
> 
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at r-project.org
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor



More information about the Bioconductor mailing list