Translate long names of proteins to short names

Question

[muk.smita]

I have a list of proteins which are identified as following:

sp|P20930|FILA_HUMAN Filaggrin OS=Homo sapiens OX=9606 GN=FLG PE=1 SV=3
sp|Q5D862|FILA2_HUMAN Filaggrin-2 OS=Homo sapiens OX=9606 GN=FLG2 PE=1 SV=1
sp|P29508|SPB3_HUMAN Serpin B3 OS=Homo sapiens OX=9606 GN=SERPINB3 PE=1 SV=2
sp|Q08188|TGM3_HUMAN Protein-glutamine gamma-glutamyltransferase E OS=Homo sapiens OX=9606 GN=TGM3 PE=1 SV=4
sp|P31025|LCN1_HUMAN Lipocalin-1 OS=Homo sapiens OX=9606 GN=LCN1 PE=1 SV=1
sp|P62805|H4_HUMAN Histone H4 OS=Homo sapiens OX=9606 GN=HIST1H4A PE=1 SV=2

Can I translate these identifiers in a more manageable form?

[Benn]

What is for you 'more manageable'? Would that be the uniprot name (e.g., P20930), or Full name (e.g., Filaggrin), or Gene Symbol (FLG)? Be more specific please.

[muk.smita]

I would like to know how I can shorten the protein identity to Full name and also gene symbol.

Thank you

[Benn]

You can use awk to extract these.

For full name something like this will work.

cat file.txt | awk 'BEGIN { FS="HUMAN " } { print $2 }' | awk '{ FS=" OS=" } { print $1 }'

For the gene symbols something like this.

cat file.txt | awk 'BEGIN { FS="GN=" } { print $2 }' | awk '{ FS=" PE=" } { print $1 }'

Answer 1

[Nicolas Rosewick]

use cut to extract the second field in your file i.e the UNIPROT id : P20930, Q5D862, etc...

cut -d "|" -f 2

-d defines the separator. Here "|"

-f defines which field you select. Here the 2nd one