Is it just me or is it that the NCBI has changed something drastic with its search engine functionality.
The problem: Simple searches seem to yield very incomplete information
The solution : Is there anyone out there
I have recently started using the slightly advanced features of the Entrez search utility at the NCBI. The NCBI education utilities do a very good job of explaining how to construct detailed searches for any biological information from its massive collection of heavily inter-linked databases.
What I set out to do was simple , a search that I have performed repeatedly in the last several months.
Identifying all chloride channel proteins with TrkA_C domains.
The keywords I used were a simple combination of “chloride” “voltage” and “TrkA” combined using the “History” option of Entrez.
In the past I was delighted to find at-least 90 sequences which I proceeded to save as an “TinyseqXML” file and then import into Excel for organization and downstream processing.
But last week I ran into a strange problem. No matter what I tried , I could not get back the earlier matches . Each of those entries were still in the database. Further the same search strategy yielded only two matches which were not int the original set of 90. Whats more disconcerting was that some matches contained entries like “p53 human tumour suppressor” which we know (and hope) has nothing to do with a chloride channel.
In my opinion something seems terribly wrong here.
Frantic emails to user support at NCBI have only gotten a response saying that ” we dont know why you cannot reproduce your search result” but use the MyNCBI option to save queries for future reproducibility.
I really don’t want to air my grievances in public but it is a tad bit disconcerting when a well tuned search engine like the NCBI starts behaving funny.