how to make full use of barcode on both directions

[刘鹏飞]

Dear all,
         I have a set of illumina 2*300 reads, after merge, I want to do the demultiplexing step. Reads were dual indexed, with 8bp barcode on each end of the reads. I used the extract_barcodes.py as following:

$ extract_barcodes.py -f inseqs_R1.fastq -c barcode_paired_stitched --bc1_len 8 --bc2_len 8 -o processed_seqs

in extract_barcodes.py, there are several options like :

--rev_comp_bc1

Reverse complement barcode 1 before writing [default: False]

--rev_comp_bc2

Reverse complement barcode 2 before writing [default: False]

which could do the reverse com of the barcode before writing.

consider the barcode in the 3'-end of the reads in fastq file, do I need to include the --rev_comp_bc2 or just use the default

thanks!

[TonyWalters]

Hello,

That depends upon the orientation of the read (I know it's a flippant answer, but the barcodes may have been recorded in 5'->3' or 3'->5' in your data) -you might have to try it with defaults first and see if the reads demultiplex with a reasonable count of sequences per sample, and if they don't, try again and reverse complement the reverse reads.

[刘鹏飞]

Thanks, TonyWalters, I got the point. Let me try!

[刘鹏飞]

Dear all,
        concerning the use of dual barcode on each end, there several other issues I am not sure when using qiime to split library:

purpose
2*300bp Miseq pair end data, with 8bp barcode on both end, after merge, I want to do the split step

1, mapping file preparation
in the barcode column, should the barcode be 8bp or be 16bp (8+8), with the barcode on both side being list as the formate of extract_barcodes.py output with barcode_paired_stitched option?

2, how to write the direction of barcode in the mapping file
if I need to use 8+8, which forward primer+reverse primer in the mapping file, need the reverse primer in the direction of 5'-->3' or reverse complemented?

3, will split_libraries_fastq.py take the dual barcode into account?
I did not see any option of split_libraries_fastq.py related to the use of dual barcode, although extract_barcodes.py has the option  barcode_paired_stitched. Am I missing something.

Thanks!

[TonyWalters]

Hello,

1. Use 16bp (8+8) of the combined forward + reverse barcode.

2. Use forward + reverse. split_libraries_fastq.py will not try to handle different orientations of your barcodes. If you do indeed have random orientations you will have to try all 4 combinations of the forward + reverse, forward (reverse complement) + reverse, forward + reverse (reverse complemented), forward (reverse complemented) + reverse (reverse complemented).
To handle the orientation of the reads, did you use the --attempt_read_reorientation option with extract_barcodes.py? If not, I'd suggest going back and running that. See http://qiime.org/scripts/extract_barcodes.html for more details about the requirements in the mapping file.

3. No, split_libraries_fastq.py expects the barcode to be a single barcode in the barcodes file specified by the -b parameter.