Large Discrepancies in Different De-multiplexing programs - help request
Hello,
I am working with Illumina PE reads from a MiSeq run and am having trouble correctly de-multiplexing them.
The fastq files generated from the run have very few reads, and I had a large (>4Gb) Undetermined file. I investigated the DemultiplexSummary text file provided with the run and found that the top 30 indexes were in fact my indexes, but for some reason(??) were not correctly filtered to the demultiplexed fastq files provided.
I then used the FASTX Barcode splitter (part of the FASTX toolkit - Hannon Lab), allowing a 1 nucleotide mismatch on the Undetermined file and recovered more reads - but still not anywhere near the number stated in the DemultiplexSummary file.
Finally, I inputted the raw data into Geneious, trimmed the reads of adapters using the BBDUK plugin, and again searched for the indexes allowing a 1 nucleotide mismatch. This found the most reads but still only approximately half of what the DemultiplexSummary states should be there.
Exact Data:
Sample 269320
Barcode CAAAAG/CTTTTG
DemultiplexSummary: 1,191,171 reads
Fastq file provided: 1, 428 reads
FASTX Barcode Splitter (reads found in Undetermined): 118,031 reads
Geneious (reads found in Undetermined after adapter trimming): 574, 498
My questions are as follows:
1 - what could be the reason for the discrepancies between programs?
2 - Perhaps my understanding of orientation is incorrect and I should be looking for the reverse complement of the barcode? For instance, my barcode was CAAAAG. I would expect to find this on R1 reads as the first 6 barcodes if the adapter was properly trimmed and I would not expect to find it at all on R2? Or am I misunderstanding the orientation of the adapter-barcode-read on R1 and R2? When I search for reads with CAAAAG I find significantly less (in all programs) than when I search for CTTTTG so if anyone could help my improve my comprehension of this I would be extremely grateful.
3 - Most importantly - How do I recover these reads?!?!
I am working with Illumina PE reads from a MiSeq run and am having trouble correctly de-multiplexing them.
The fastq files generated from the run have very few reads, and I had a large (>4Gb) Undetermined file. I investigated the DemultiplexSummary text file provided with the run and found that the top 30 indexes were in fact my indexes, but for some reason(??) were not correctly filtered to the demultiplexed fastq files provided.
I then used the FASTX Barcode splitter (part of the FASTX toolkit - Hannon Lab), allowing a 1 nucleotide mismatch on the Undetermined file and recovered more reads - but still not anywhere near the number stated in the DemultiplexSummary file.
Finally, I inputted the raw data into Geneious, trimmed the reads of adapters using the BBDUK plugin, and again searched for the indexes allowing a 1 nucleotide mismatch. This found the most reads but still only approximately half of what the DemultiplexSummary states should be there.
Exact Data:
Sample 269320
Barcode CAAAAG/CTTTTG
DemultiplexSummary: 1,191,171 reads
Fastq file provided: 1, 428 reads
FASTX Barcode Splitter (reads found in Undetermined): 118,031 reads
Geneious (reads found in Undetermined after adapter trimming): 574, 498
My questions are as follows:
1 - what could be the reason for the discrepancies between programs?
2 - Perhaps my understanding of orientation is incorrect and I should be looking for the reverse complement of the barcode? For instance, my barcode was CAAAAG. I would expect to find this on R1 reads as the first 6 barcodes if the adapter was properly trimmed and I would not expect to find it at all on R2? Or am I misunderstanding the orientation of the adapter-barcode-read on R1 and R2? When I search for reads with CAAAAG I find significantly less (in all programs) than when I search for CTTTTG so if anyone could help my improve my comprehension of this I would be extremely grateful.
3 - Most importantly - How do I recover these reads?!?!