Demultiplex Illumina reads

[newBioinfo]

Hi Everyone,
I am kind of stuck with my Illumina data, I want to remove the barcodes from my reads. My read file looks like this
@HWI-ST1035:115:C0RG7ACXX:5:1101:1216:2040 1:N:0:
NACAGAGGATGCAAGCGTTATCCGGAATGATTGGGCGTAAAGCGTCTGNNNGNNNNNNNNNNNNNNNNNNNNNNNNNNNN
NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN
+
#1=DDDFFCADHHIIIIIIIIGIIIIIIIGIIIIIIIFHIIIIICFHH################################
#######################################################################
and my barcode files look like this:
@HWI-ST1035:115:C0RG7ACXX:5:1101:1216:2040 2:N:0:
NNNNNANACACA
+
############
When I am using fastx toolkit to trim barcodes I am getting error.
The command I am using is:
cat lane5_NoIndex_L005_R1_001.fastq | /u2/software/fastx/fastx_toolkit-0.0.13.2/bin/fastx_barcode_splitter.pl --bcfile lane5_NoIndex_L005_R2_001.fastq --bol --prefix x --suffix ".fastq"
The error I am getting is:
Error: bad barcode value (2:N:0 at barcode file (lane5_NoIndex_L005_R2_001.fastq) line 1
The reason I think is beacuse of 2:N:0 in the barcode header and 1:N:0 in the reads header.
I am not sure how to rectify this, please if anyone has any idea could you please help me.

Thanks!!!!

[JackieBadger]

Use Trimmomatic... much more versatile and mate-pair aware. Or just use the trim function in text manipulation in Galaxy.

[sklages]

Why don't you ask your sequencing provider to demultiplex the data for you? It is not more work for them, as it is part of the fastq processing; you just need to provide some kind of sample description, a samplesheet.

All our customers are quite happy that this work has already been done when they get their data :-)

Sven

[JackieBadger]

BEWARE! You must always have the ability to check any processing a provider does for you... for e.g. the trimming script on the MiSeq software should be avoided as it is VERY promiscuous...even in the "remedied" latest update. Also...I have come across significant errors in the MiSeq de-multiplexing.
I would do everything with a program where you set the parameters and know what is going in and what should come out.

[newBioinfo]

Use Trimmomatic... much more versatile and mate-pair aware. Or just use the trim function in text manipulation in Galaxy.

Thanks jackieBadger,
I will try it!

[newBioinfo]

BEWARE! You must always have the ability to check any processing a provider does for you... for e.g. the trimming script on the MiSeq software should be avoided as it is VERY promiscuous...even in the "remedied" latest update. Also...I have come across significant errors in the MiSeq de-multiplexing.
I would do everything with a program where you set the parameters and know what is going in and what should come out.

Thanks JackieBadger,
I am planning to use QIME, so hopefully I will not encounter such issues.



Thanks for the help!!!

[newBioinfo]

Why don't you ask your sequencing provider to demultiplex the data for you? It is not more work for them, as it is part of the fastq processing; you just need to provide some kind of sample description, a samplesheet.

All our customers are quite happy that this work has already been done when they get their data :-)

Sven

Thanks Seven,
Ii would be good if they demultiplex the data before sending, but in my case it is not.

[sklages]

BEWARE! You must always have the ability to check any processing a provider does for you... for e.g. the trimming script on the MiSeq software should be avoided as it is VERY promiscuous...even in the "remedied" latest update. Also...I have come across significant errors in the MiSeq de-multiplexing.
I would do everything with a program where you set the parameters and know what is going in and what should come out.

hhm, it's pretty easy to check what the provider does if you also have the "Undetermined_indices" data files. MiSeq is another thing ... the trimming issue is known and should not be used (currently). You could also ask for some (demultiplexing) stats, to see if the results are "good" or as expected.

If you don't trust in your sequence provider at all, you should look for another one ;-)

What "significant errors" did you encounter in the MiSeq demultiplexing?
We are not plexing Miseq libs, so I am just curious :-)

Sven

[NextGenSeq]

Are you sure they did an index read? The title of your files say No_Index. The only time we ever get data titled like this is if a index read wasn't done.

[sklages]

Are you sure they did an index read? The title of your files say No_Index. The only time we ever get data titled like this is if a index read wasn't done.

No, you'll always get that naming when there is not index specified in the samplesheet for that run (irrespective if there was run an index read).
You cannot safely deduce from the naming wether there has been run an index read or not (at least for the "_NoIndex_" case)..

Sven

[GenoMax]

No, you'll always get that naming when there is not index specified in the samplesheet for that run (irrespective if there was run an index read).
You cannot safely deduce from the naming wether there has been run an index read or not (at least for the "_NoIndex_" case)..

Sven

Let us hope that is the case. If the facility did not run this as a multiplex sample then OP is out of luck. This run will have to be repeated.

[sklages]

Let us hope that is the case. If the facility did not run this as a multiplex sample then OP is out of luck. This run will have to be repeated.

Sure, you are absolutely right.

This problem might arise if the customer doesn't mention any indices that need to be demultiplexed in their "order" (however this order looks like), maybe assuming that this is not relevant for the sequencing run itself but for the post-processing only.
... and the sequencing core organizes their FCs with respect to read length and MP/no MP ...

We had a similar post a while ago, where the OP has hand-written a little note on the "order sheet" and as a result the sequencing didn't recognize it as "please do an index read, as my libraries have indices" ...

Sven

[LVAndrews]

Thanks JackieBadger,
I am planning to use QIME, so hopefully I will not encounter such issues.



Thanks for the help!!!

If you are using QIIME, then you will have the option to remove unwanted sequences (such as barcodes) during the split_libraries_fastq.py step. See http://qiime.org/scripts/split_libraries_fastq.html for more information.

[LVAndrews]

If you are using QIIME, then you will have the option to remove unwanted sequences (such as barcodes) during the split_libraries_fastq.py step. See http://qiime.org/scripts/split_libraries_fastq.html for more information.

Oops for this application you want split_libraries.py script instead.