This is a test version of Biostars. For the public version, visit https://www.biostars.org.
NCBI esearch results inconsistent with web-based search results?

I'm trying to collect (and then download) all transcriptomes and genomes associated with txid7604.

The query I am using is:

((txid7604[Organism:exp]) AND ( "tsa master"[Properties] OR "wgs master"[Properties] ))

And the database I am looking at is nuccore.

When I search via the web interface I get four matches: ncbinuccore_web

However, when I query via esearch 16.2 (installed via bioconda), I get 81468 matches?

> esearch -db nuccore -query "((txid7604[Organism:exp]) AND ("tsa master"[Properties] OR "wgs master"[Properties]))"
<ENTREZ_DIRECT>
  <Db>nuccore</Db>
  <WebEnv>MCID_638e1b0b374c2370e333fddb</WebEnv>
  <QueryKey>1</QueryKey>
  <Count>81468</Count>
  <Step>1</Step>
</ENTREZ_DIRECT>

What is the source of this discrepancy?

ncbi eutils
• 1,532 views • link
updated by GenoMax GenoMax • written by Dunois • Dunois

2 answers

This is a quoting issue, try to replace the outer double quotes with single quotes.

esearch -db nuccore -query '((txid7604[Organism:exp]) AND ("tsa master"[Properties] OR "wgs master"[Properties]))'
<ENTREZ_DIRECT>
  <Db>nuccore</Db>
  <WebEnv>MCID_638e1db1b926d65727335953</WebEnv>
  <QueryKey>1</QueryKey>
  <Count>4</Count>
  <Step>1</Step>
</ENTREZ_DIRECT>

In principle, you should also have gotten a warning like this for your original query:

esearch -db nuccore -query "((txid7604[Organism:exp]) AND ("tsa master"[Properties] OR "wgs master"[Properties]))"
Entrez Direct does not support positional arguments.
Please remember to quote parameter values containing
whitespace or shell metacharacters.
• 0 views • link
written by Michael Dondrup Michael Dondrup

I see. I actually managed to find out how to circumvent this.

esearch (and I suppose all its siblings) make a distinction between tsa-master and tsa master and wgs-master and wgs master with double quotes. Adding the hyphen resolves the discrepancy.

No hyphenation -> incorrect count:

> esearch -db nuccore -query "((txid7604[Organism:exp]) AND ("tsa master"[Properties] OR "wgs master"[Properties]))"
<ENTREZ_DIRECT>
  <Db>nuccore</Db>
  <WebEnv>MCID_638e1ebf93dc63259250e57f</WebEnv>
  <QueryKey>1</QueryKey>
  <Count>81468</Count>
  <Step>1</Step>
</ENTREZ_DIRECT>

With hyphenation, correct result:

> esearch -db nuccore -query "((txid7604[Organism:exp]) AND ("tsa-master"[Properties] OR "wgs-master"[Properties]))"
<ENTREZ_DIRECT>
  <Db>nuccore</Db>
  <WebEnv>MCID_638e1ec7b18a76276123d0dc</WebEnv>
  <QueryKey>1</QueryKey>
  <Count>4</Count>
  <Step>1</Step>
</ENTREZ_DIRECT>
• 0 views • link
written by Dunois • Dunois 
 esearch -db nuccore -query '((txid7604[Organism:exp]) AND ("tsa master"[Properties] OR "wgs master"[Properties]))'

Quotes matter, shell is going to split your command as separate parameters if not quoted properly. Try wrapping your query with single quotes.

https://colab.research.google.com/drive/15xJPdwsNqeU3u-3h-kmNTR2XEaeVggzM?usp=sharing

• 0 views • link
updated by GenoMax GenoMax • written by barslmn • barslmn

Log in to answer this question.

More posts like this