Using entrez-direct inside a for loop in bash

Question

[Lille My]

I'm trying to retrieve the genome from which a series of proteins are derived. there is more than one assembly for each protein, so I need to create a file where they are linked.

I use the following:

for id in `cat gi-list-file`; do
elink -target nuccore -db protein -id $id |
elink -target assembly |
esummary  |
 xtract -pattern AssemblyAccession -element AssemblyAccession
done

the first result I get is the assembly accession, but the second result is the following error message:

Retrying elink, step 2: callMLink: Error reading an UID blob, ,CNCHistory::ReadIdListBuf, result (false) error, blobid=empty

Any ideas on what the problem is?

[Joe]

what happens if you run that accession directly outside of the loop?

Can you also show us some examples of accessions which work, and some that don't?

[Christina]

Hi! I tried to do something similar, but it doesn't work. I have a list of Pubmed IDs and I want to retrieve their abstracts.

for i in `cat only_retrieved_pubmedIDs.csv` ; do echo $i; efetch -input $i -db pubmed -format abstract > $i.txt ; done

One file per Pubmed ID is produced but they are all empty.

[GenoMax]

While your questions is unrelated to the original thread you should do the following (one PMID per line in input file):

$ for i in `cat id_file`; do efetch -db  pubmed -id ${i} -format abstract; done

Answer 1

[GenoMax]

Following works for me.

$ more names.txt
WP_043107373
WP_000617546.1
WP_000906486.1
WP_001096206.1
WP_001386830.1

Loop used for the lookups

$ for i in `cat names.txt`; do echo $i; elink -target nuccore -db protein -id $i |elink -target assembly|esummary |xtract -pattern AssemblyAccession -element AssemblyAccession > $i.txt; done

This should produce one file per input Accession number.

With a non-existent accession number an error will be generated and result in an empty file for that accession. Loop should continue for rest of the accession.

WP_031373
ERROR in link output: BLOB ID IS NOT IMPLEMENTED

Actual error message is larger, truncated for display.