merge gbk files

[krispy]

Hi all, this might sound stupid but I would like to know how to combine multiple genbank files into one continuous file. I tried using <cat> to combine 4 gbk files from 4 chromosomes but when I try to view it on Artemis, only the first can be displayed. I found a thread where people suggested UGENE but I don't know why the software only put the annotation from 4 files side by side, but did not shift the nucleotide and amino sequences accordingly and it all appeared to be all NNNNN or XXXXX. I would like to view all the 4 chromosomes as one full genome in Artemis, so....if anyone can help me with this. Thanks and have a nice day!!!

[maubp]

Concatenating GenBank files to make a multi-record GenBank file should work fine - the question is how to get Artemis to display it nicely

From memory last time I did this I had to merge the GenBank files into a single dummy record, for which I added an NNNNNNNNN spacer between the records so this break was shown in Artemis visually. I did that with Biopython if that is an interesting solution for you.

[JohnN]

union from the EMBOSS suite will work, eg.

$ union -sequence FileA.gbk -sformat genbank -outseq FileB.gbk -osformat genbank -feature Y -auto

[Roy]

Originally Posted by maubp

I did that with Biopython if that is an interesting solution for you.

If BioPerl is more your thing then look at Bio::SeqUtils->cat

[zhangju]

Originally Posted by maubp

Concatenating GenBank files to make a multi-record GenBank file should work fine - the question is how to get Artemis to display it nicely

From memory last time I did this I had to merge the GenBank files into a single dummy record, for which I added an NNNNNNNNN spacer between the records so this break was shown in Artemis visually. I did that with Biopython if that is an interesting solution for you.

Would you be able to give some details how to use biopython to merge multiple genbank files into a single sequence with Ns and keep all the annotations with adjusted coordinates?

Thanks,

Justin

[zhangju]

Originally Posted by Roy

If BioPerl is more your thing then look at Bio::SeqUtils->cat

Could number of Ns be added between two sequences at merging with Bioperl?

Thanks,

Justin

[maubp]

Something like this, where you should replace "N"*50 with your desired spacer (this would be 50 N characters), and generalise for as many records as you have:

from Bio import SeqIO
recordA = SeqIO.read("fileA.gbk", "genbank")
recordB = SeqIO.read("fileB.gbk", "genbank")
combined_record = recordA + ("N" * 50) + recordB
SeqIO.write(combined_record, "combined.gbk", "genbank")

[Giffredo]

Hi,
the python script doesn't work on my genbank files.

"ValueError: Need a Nucleotide or Protein alphabet"

Anyway.. If I have several genbank files can I use the function:

in_genbank = open(sys.argv[2], 'r')
out_genbank = open(sys.argv[3], 'w')

instead of:

recordA = SeqIO.read("fileA.gbk", "genbank")
recordB = SeqIO.read("fileB.gbk", "genbank")
SeqIO.write(combined_record, "combined.gbk", "genbank")

where sys.arg[2] is: *.gbk

$python gbk_merge.py *.gbk out.gbk

[maubp]

Normally the first line (the LOCUS line) of a GenBank file says if the record is DNA or protein - and Biopython uses that information. If this failed I would expect this error on trying to save as GenBank format:

"ValueError: Need a Nucleotide or Protein alphabet"

Can you show us the start of your GenBank records? e.g. at the Linux terminal prompt try:

$ head fileA.gbk

Once you have this working to merge two GenBank records with a spacer we can generalise this to merging many.

[Giffredo]

This is the head:

$ head fileA.gbk
LOCUS DG_Contig_04
DEFINITION seqnum=1;seqlen=498669;seqhdr="DG_final_Contig_04";version=Prodigal.v2.50;run_type=Single;model="Ab initio";gc_cont=46.95;transl_table=11;uses_sd=1
FEATURES Location/Qualifiers
gene complement(3..413)
/ID="LPDG_0036"
/locus_tag="LPDG_0036"
/note="ID=1_1;partial=10;start_type=ATG;rbs_motif=AGxAGG/AGGxGG;rbs_spacer=5-10bp;score=21.71;cscore=5.19;sscore=16.52;rscore=10.95;uscore=2.29;tscore=3.94"
CDS complement(3..413)
/ID="LPDG_0036"
/locus_tag="LPDG_0036"

$head fileB.gbk
LOCUS DG_Contig_05
DEFINITION seqnum=1;seqlen=45471;seqhdr="DG_final_Contig_05";version=Prodigal.v2.50;run_type=Single;model="Ab initio";gc_cont=46.03;transl_table=11;uses_sd=1
FEATURES Location/Qualifiers
gene 3..176
/ID="LPDG_0520"
/locus_tag="LPDG_0520"
/note="ID=1_1;partial=10;start_type=Edge;rbs_motif=None;rbs_spacer=None;score=1.94;cscore=-0.73;sscore=2.68;rscore=0.00;uscore=3.18;tscore=0.00"
CDS 3..176
/ID="LPDG_0520"
/locus_tag="LPDG_0520"

[maubp]

You need to use the forums's [ code ] and [ /code ] tags (without spaces) in order to preserve the white space - but this confirms you don't have proper GenBank files.

Biopython should be giving you a warning, but the workaround is to tell the parser these are DNA sequences. Does this work for you with two files?

from Bio.Alphabet import generic_dna
from Bio import SeqIO
recordA = SeqIO.read("fileA.gbk", "genbank", generic_dna)
recordB = SeqIO.read("fileB.gbk", "genbank", generic_dna)
combined_record = recordA + ("N" * 50) + recordB
SeqIO.write(combined_record, "combined.gbk", "genbank")

[Giffredo]

now I receive this error:

ValueError: Specified alphabet DNAAlphabet() clashes with that determined from the file, Alphabet()

I think I have not Bio.Alphabet...

[maubp]

Can you email me the two input files so I can test this locally (on GoogleMail - I'll message you via the forum)? Also please confirm which version of Biopython you have.

[maubp]

Thanks for emailing the files. Hmm. Where did this "GenBank" files come from? That would be useful for us to know (e.g. for other people using the same tool), but right now my suggestion for setting the alphabet does not work.

Workaround (tested):

from Bio.Alphabet import generic_dna
from Bio import SeqIO
recordA = SeqIO.read("fileA.gbk", "genbank")
recordA.seq.alphabet = generic_dna
recordB = SeqIO.read("fileB.gbk", "genbank")
recordB.seq.alphabet = generic_dna
combined_record = recordA + ("N" * 50) + recordB
SeqIO.write(combined_record, "combined.gbk", "genbank")

[maubp]

Version which can be used on multiple files. In order to keep the interface simple this just takes input filenames and writes the record to stdout, e.g.

$ python merge_gbk.py fileA.gbk fileB.gbk > combined.gbk
Loading fileA.gbk
Loading fileB.gbk
Merging into one record

Or, you could use it this, but be careful using *.gbk as the input if your output file will also be in the same directory

$ python merge_gbk.py input_folder/*.gbk > output_folder/combined.gbk

The script also works on Python 3,

#!/usr/bin/env python
# Copyright 2016 Peter Cock, James Hutton Institute
# This example Biopython script is released for free re-use under
# the MIT license https://opensource.org/licenses/MIT
# For backgound see http://seqanswers.com/forums/showthread.php?t=28924
import sys
from Bio.Alphabet import generic_dna
from Bio import SeqIO

if len(sys.argv) < 2:
    print("Usage: Takes as arguments filenames for input DNA GenBank files,")
    print("outputs merged record to stdout using 50bp runs of N as linker.")
    print("")
    print("python merge_gbk.py input*.gbk > combined.gbk")

# argv[0] = python script
# argv[1] = first record
# argv[2] onwards = subsequence records
filename = sys.argv[1]
sys.stderr.write("Loading %s\n" % filename)
combined_record = SeqIO.read(filename, "genbank")
# Forcing DNA alphabet in case of improper input files
combined_record.seq.alphabet = generic_dna
for filename in sys.argv[2:]:
    sys.stderr.write("Loading %s\n" % filename)
    record = SeqIO.read(filename, "genbank")
    record.seq.alphabet = generic_dna
    combined_record = combined_record + ("N" * 50) + record
sys.stderr.write("Merging into one record\n")
count = SeqIO.write(combined_record, sys.stdout, "genbank")
assert count == 1, count