make fasta sequences names short

Question

[radha.jg]

Hi,

I'm a newbie so please be patient.

I have a fasta file like this:

>gi|820716087|gb|AKG62099.1| eIF-2 alpha kinase [Leishmania donovani]
MAKKKNECHSCRLVQAYNTCENDEIKDEIDIIVNTYENVRVSGKSAAHYRVLVPLTSESHPSRRVTLEIR
VVPGYPYVVPAINLLFPPGLQPGCEGTLSEYEVKQMAKEVLNNIQPCLPSGMPCMMQIVSTVASIVECSI
DPPSQQQNGKAQGEPKVLSAGQSSSLTPVPLKAKEALKLSLFAFHLLKKCCHMKNPESNEEAASNFDWLV
KYLLDSVRIFPEAARSFFPWNGISSSRAFAANIESALALPPDQQGLPKWLWEDEGRNPRIQQGSEGRYRN
>gi|820957452|pdb|4WZH|B Chain B, Dihydroorotate Dehydrogenase From Leishmania Viannia Braziliensis
MGSSHHHHHHSSGLVPRGSHMASMTGGGQMGRGSMSLQVGILGNTFANPFMNAAGVMCSTEEELAAMTES
TSGSLITKSCTPALREGNPAPRYYTLPLGSINSMGLPNKGFDFYLAYSARHHDYSRKPLFISISGFSAEE
NAEMCKRLAPVAAEKGVILELNLSCPNVPGKPQVAYDFDAMRRYLAAISEAYPHPFGVKMPPYFDFAHFD
AAAEILNQFPKVQFITCINSIGNGLVIDVETESVVIKPKQGFGGLGGRYVFPTALANVNAFYRRCPGKLI
FGCGGVYTGEDAFLHVLAGASMVQVGTALHEEGAAIFERLTAELLDVMAKKGYKALDEFRGKVKAMD

How do I transform the names to have something like this:

>gb_AKG62099.1
MAKKKNECHSCRLVQAYNTCENDEIKDEIDIIVNTYENVRVSGKSAAHYRVLVPLTSESHPSRRVTLEIR
VVPGYPYVVPAINLLFPPGLQPGCEGTLSEYEVKQMAKEVLNNIQPCLPSGMPCMMQIVSTVASIVECSI
DPPSQQQNGKAQGEPKVLSAGQSSSLTPVPLKAKEALKLSLFAFHLLKKCCHMKNPESNEEAASNFDWLV
KYLLDSVRIFPEAARSFFPWNGISSSRAFAANIESALALPPDQQGLPKWLWEDEGRNPRIQQGSEGRYRN
>pdb_4WZH
MGSSHHHHHHSSGLVPRGSHMASMTGGGQMGRGSMSLQVGILGNTFANPFMNAAGVMCSTEEELAAMTES
TSGSLITKSCTPALREGNPAPRYYTLPLGSINSMGLPNKGFDFYLAYSARHHDYSRKPLFISISGFSAEE
NAEMCKRLAPVAAEKGVILELNLSCPNVPGKPQVAYDFDAMRRYLAAISEAYPHPFGVKMPPYFDFAHFD
AAAEILNQFPKVQFITCINSIGNGLVIDVETESVVIKPKQGFGGLGGRYVFPTALANVNAFYRRCPGKLI
FGCGGVYTGEDAFLHVLAGASMVQVGTALHEEGAAIFERLTAELLDVMAKKGYKALDEFRGKVKAMD

The idea is to have just the genebank id, or, if it's not in the name, one of the ids and where is it from

Saludos :)

Answer 1

[Devon Ryan]

awk 'BEGIN{FS="|"}{if(NF>1) {printf(">%s_%s\n", $3, $4)}else{print $0}}' foo.fa > fixed.fa

[radha.jg]

Excellent. Thank you very much. I seriously need to learn how to program in awk.

[radha.jg]

In case of sp_, how do I use an unique identifier like gi? I assume that $1 and $2 will do

[Devon Ryan]

You could check the number of fields (NF) in a more elaborate way and use $1 and $2 as needed.

[wang-yanfang]

Hey Devon Ryan,

Could you please help me with some modification of your code for my problem?

I also want to shorten the fasta file sequence header, which looks like this:

>lcl|VSMA01000001.1_prot_KAB5584702.1_1 [locus_tag=GE09DRAFT_1165795] [db_xref=InterPro:IPR002198,JGIDB:Conioc1_1165795] [protein=tetrahydroxynaphthalene reductase] [protein_id=KAB5584702.1] [location=join(1826..1931,1988..2458,2736..2863,2927..3064)] [gbkey=CDS]
MPGLTTNTGKYDQIPGPLGLASASLEGKVALVTGAGRGIGREMAQELGRRGAKVIVNYANSQESAEEVVQAIKKSGSDAA
SIKANVSDVDQIVRMFDEAVKVFGKLDIVCSNSGVVSFGHVKDVTPEEFDRVFNINTRGQFFVAREAYKHLEVGGRLILM
GSITGQAKGVPKHAVYSGSKGTIETFVRCMAIDFGDKKITVNAVAPGGIKTDMYHAVCREYIPNGINLTDDEVDEYACTW
SPLHRVGLPIDIARVVCFLASQDGEWINGKVLGIDGAACM
>lcl|VSMA01000001.1_prot_KAB5584705.1_4 [locus_tag=GE09DRAFT_52] [db_xref=InterPro:IPR010730,JGIDB:Conioc1_52] [protein=heterokaryon incompatibility protein-domain-containing protein] [protein_id=KAB5584705.1] [location=10796..11233] [gbkey=CDS]
MPTRLLEIDPQANSRHIRLVSDTGILLKERYAALSHCWGKSPTNTTTKAVFVSHTQGIDILSLSKTFQHTIFVTRELGIR
YLWIDSLCIIQDDEDDWKREAENMADVFANAFVTIAASASTDGDGGLFYPRALETERSGTVRWTI

And I want the header to be this:

>GE09DRAFT_1165795
>GE09DRAFT_52

And I tried your code with

awk 'BEGIN{FS=" "}{if(NF>1) {printf(">%s\n", $2)}else{print $0}}' in.fasta > out.fasta

It gave me the results:

>[locus_tag=GE09DRAFT_1165795]
..

SO How can I cut the [locus_tag= and ]?

I would be really appreciated for any help.

Thanks!
Yanfang

[wang-yanfang]

I think I figured this out by adapting two codes together.

awk 'BEGIN{FS=" "}{if(NF>1) {split($2,a,"="); split(a[2],b,"]"); printf(">%s\n",b[1])}else{print $0}}' in.fasta > out.fasta

Thanks all the help!

Yanfang