Mapping refseq transcripts to encoded proteins (NM_ to NP_)

Question

[Endre Bakken Stovner]

I'd like to find the protein encoded by my refgene transcripts.

I have the refGene.txt file where each line looks like the following:

138    NM_016166    chr15    +    68346571    68480404    68346664    68480173    14    68346571,68378643,68434283,68434627,68438153,68438903,68445927,68457068,68466069,68467974,68468811,68473549,68475967,68479879,    68346688,68379088,68434368,68434675,68438244,68439038,68446033,68457142,68466230,68468105,68468992,68473692,68476005,68480404,    0    PIAS1    cmpl    cmpl    0,0,1,2,2,0,0,1,0,2,1,2,1,0,

I'd like to know how I can get the NP_XXXX name for the transcript NM_016166.

Preferably, I'd like to just get a text file mapping the two in some way, but I'll accept any answer that does this automatically, e.g. with BioPython (having to look it up by hand in a browser or some such doesn't cut it - I need to do this for 50K transcripts).

I'm working with hg38, but I'm guessing the procedure is the same for all major genome versions, so I did not specify to make the Q as general as possible.

Answer 1

[5heikki]

Here's one way with Entrez Direct:

IFS=$'\n'; for next in $(cut -f2 refGene.txt); do NP=$(esearch -db nuccore -query $next | elink -target protein | efetch -format docsum | xtract -element Caption); echo "$next $NP"; done

[Endre Bakken Stovner]

Thanks, I need to install some tools so will get back to you. Upvote.

Ps. other answers still welcome. Is there a file that contains these mappings somewhere?

[Endre Bakken Stovner]

Your previous, non-while version was better. This one includes an error (and only reads the first line anyways, see this post on StackOverflow.

[5heikki]

My, bad. I edited the answer again.

Answer 2

[Endre Bakken Stovner]

Hello myself from the past! After having done similar tasks umpteen gazillion times you realized you should make them easier for yourself. Therefore you wrote biomartian (pip install biomartian:

echo "138 NM_016166 chr15 + 68346571 68480404 68346664 68480173 14" | sed 's/ /\t/g' | biomartian --noheader -c 1 -i refseq_mrna -o refseq_peptide -
0    1    2    3    4    5    6    7    8    refseq_peptide
138    NM_016166    chr15    +    68346571    68480404    68346664    68480173    14    NP_057250

(The code before the biomartian call is just used to replace spaces with tabs).

See https://github.com/endrebak/biomartian for more

[Malcolm.Cook]

this example does not work in my hands. The refseq_peptide returns blank.

Answer 3

[Endre Bakken Stovner]

The UCSC table to map transcripts to proteins is called refLink.

You can get it as a plain-text file by going to the UCSC table browser, selecting your genome of interest and then setting the group to "all tables". Next choose the table refLink, enter a filename and then press the "get output"-button.

Example output:

#name product mrnaAcc protAcc geneName prodName locusLinkId omimId
PAX2 paired box 2 NM_001282819 NP_001269748 133420 110035 102094402 0
MLPH melanophilin NM_001282836 NP_001269765 65090 92499 102093006 0

[Li Hsing]

Great! Your answer help me a lot!