Demultiplexing 454

[inesdesantiago]

Dear all,
I am trying to de-multiplex a 454 fasta file.

It is a public data file. In the paper (PMID:21731642) the authors say that:
"The HB4a and C5.2 Y-shaped adapters were formed by primers A and B and primers C and D, respectively (Primer A: 5′-GATCTCCCGAGTGGTCACCTGCTC-3′; Primer B: 5′-CTAGCAGCTACCACTCGGGA-3′; Primer C: 5′-GATCCCCTGAGTGGTCACCTGCTC-3′, and Primer D: 5′-CTAGCAGCTACCACTCAGGG-3′)"

So I did two runs of demultiplexing:

Run1:

$ cat barcode1.txt
HB4a_1  GATCTCCCGAGTGGTCACCTGCTC
C52_1     GATCCCCTGAGTGGTCACCTGCTC

$ cat ../HER2_Dataset_Download/SRR342054_1.fastq | fastx_barcode_splitter.pl --bcfile barcode1.txt --bol --mismatches 2 --prefix SRR342054_1_ --suffix ".fastq" 

Barcode	Count	Location
C52_1	0	SRR342054_1_C52_1.fastq
HB4a_1	0	SRR342054_1_HB4a_1.fastq
unmatched	802214	SRR342054_1_unmatched.fastq

Run2:

$ cat barcode2.txt
HB4a_2  CTAGCAGCTACCACTCGGGA
C52 _2    CTAGCAGCTACCACTCAGGG

cat SRR342054_1_unmatched.fastq | fastx_barcode_splitter.pl --bcfile barcode2.txt --bol --mismatches 2 --prefix SRR342054_2_ --suffix ".fastq"

Barcode	Count	Location
C52_2	214558	SRR342054_2_C52_2.fastq
HB4a_2	165481	SRR342054_2_HB4a_2.fastq
unmatched	422175	SRR342054_2_unmatched.fastq
total	802214

On the first run there are no reads that map those 2 sequences.
On the second run there are many unmatched reads (>50% of reads do no match any barcode)

Am I doing something wrong?
Does anyone have experience with this type of problem?

Thanks!

[inesdesantiago]

Actually, the reverse complement of the first barcode worked.
I decided to allow only 1 mismatch now

Barcode1 output
Barcode Count Location
C52_1 166119 SRR342054_1rev_C52_1.fastq
HB4a_1 134556 SRR342054_1rev_HB4a_1.fastq
unmatched 501539 SRR342054_1rev_unmatched.fastq
total 802214

Barcode 2 output
Barcode Count Location
C52_2 213282 SRR342054_2rev_C52_2.fastq
HB4a_2 135893 SRR342054_2rev_HB4a_2.fastq
unmatched 152364 SRR342054_2rev_unmatched.fastq
total 501539