Odd behaviour - NCBI Eutils

Question

[npb27]

Hi,

I'm having some weird behaviour for esearch interpreting variables from the bash command line. I have a list of NCBI protein IDs, and I want to obtain the nucleotide database ID for the associated genome/contig as well as the genome length

This works:

esearch -db protein -query "GIR74039.1" | elink -target nuccore | efetch -format docsum

This works:

i="GIR74039.1"
esearch -db protein -query $i | elink -target nuccore | efetch -format docsum

This doesn't work:

INPUT contains only the protein sequence ID shown above

cat INPUT | while read i
do
echo $i
esearch -db protein -query ${i} | elink -target nuccore | efetch -format docsum > temp
done

Here's the error I'm getting

WARNING:  FAILURE ( Mon  4 Mar 15:58:07 GMT 2024 )
 -tool edirect -edirect 16.2 -edirect_os Linux -email [e-mail]0 -usehistory y -db protein -term GIR74039.1
<ERROR>Empty term and query_key - nothing todo</ERROR>
SECOND ATTEMPT
 WARNING:  FAILURE ( Mon  4 Mar 15:58:09 GMT 2024 )
 -tool edirect -edirect 16.2 -edirect_os Linux -email [e-mail]0 -usehistory y -db protein -term GIR74039.1
<ERROR>Empty term and query_key - nothing todo</ERROR>
LAST ATTEMPT
 ERROR:  FAILURE ( Mon  4 Mar 15:58:11 GMT 2024 )
 -tool edirect -edirect 16.2 -edirect_os Linux -email [e-mail]0 -usehistory y -db protein -term GIR74039.1
<?xml version="1.0" encoding="UTF-8" ?>
<!DOCTYPE eSearchResult>
<eSearchResult>
  <ERROR>Empty term and query_key - nothing todo</ERROR>
</eSearchResult>
QUERY FAILURE
 ERROR:  WebEnv value not found in elink input
 ERROR:  Missing -db argument

(I censored my email from the output)

I've tried all combinations I can think of single/double quotes and {} around the $i variable, none help

I think the record is still being found by esearch as it's mentioned in the error. This seems like an odd error from esearch

Answer 1

[GenoMax]

Can you provide a list of ID's that are failing? Do they work if you try them individually without the while loop?

You may want to use epost method instead of the loop.

[npb27]

One ID that isn't working is shown in the question, they work individually

Many thanks, I wasn't aware of epost but this solved the issue

This now works:

head -n1 INPUT | while read i
do
echo $i
echo $i | epost -db protein  | elink -target nuccore | efetch -format docsum > temp
done

[K]

Same problem here :

esearch -db pubmed -query "(RAB5C-AS1)OR(ENSG00000267658)"

With the error :

WARNING:
FAILURE ( lun. déc.  9 12:15:32 CET 2024 )
nquire -url https://eutils.ncbi.nlm.nih.gov/entrez/eutils/ esearch.fcgi -retmax 0 -usehistory y -db pubmed -term \(RAB5C-AS1\)OR\(ENSG00000267658\) -tool edirect -edirect 14.9 -edirect_os Linux -email xxxx
<ErrorList>
  <PhraseNotFound>RAB5C-AS1</PhraseNotFound>
  <PhraseNotFound>ENSG00000267658</PhraseNotFound>
<ErrorList>

[GenoMax]

Following produces a result:

$ esearch -db pubmed -query "RAB5C antisense RNA 1"
<ENTREZ_DIRECT>
  <Db>pubmed</Db>
  <QueryKey>1</QueryKey>
  <Count>25</Count>
  <Step>1</Step>
</ENTREZ_DIRECT>

[K]

I agree, I had indeed noticed that you could access the result by modifying certain formulations. However, this doesn't answer the question "what's stopping the query from working?" because I have a very large number of queries that don't work. I can't afford to do manual checks for automatic querying ...

[GenoMax]

Only way would be to report this to NCBI help desk. They will respond but be ready to wait for a few business days.