Demultiplexing And Splitting Miseq Fastq

Question

[peris]

I got the following files from a service provider while doing MiSeq sequencing. Can someone help me in demultiplexing and splitting this files? I have used 24 independent indexes for indexing.

Read1:

@130110_A5T23:1:1101:15646:1320:1#0/1
NAGTAAGATACAGTCTATCGGGTTTAAGTTATACAACATAGTACAGTACATTCATACCTACCTCTGCAATTAAATTTGGCGGTGTCATAATGTCTTTCAGCACAACTAAAAGAAAAGTTTAAAAGTGATATAAAGTTATCTTTTGCACTT
+
#>>AABFFFFFFGGGGGGGGGGCFGHHHHHHHHHHHHHHHHGFHHHHHHHHGHGHHHHHHGHHGHHHGHHHHHHHHHEHGGGEEFGHHHHFHHHGHHHHFHHHHGGGHHHHGGHHGHHHHHHHGHBGHHHHHHHHHFHFHHHHHGHFHHH
@130110_A5T23:1:1101:17282:1320:0#0/1
NGGGAGGATGGCTTGAGCCTAGAGGGTTGAGGCTGCAGTGAGCTGTGAGCATGCCATTGCACTCCAGCCTGGGCAACAGAGTGAGGCTCTGTCTCAAAAAAAAAAAAAAAAAAAAAAGAAATAAAATATAAGGGACATAAAATAAAGCAA
+
#>>AAA1A1B>F1A1FAFGCG0GBFEAAEHGBFCFFHHA1FEGGFFGGFFHHHGHHHHGHHHHHHHHHGHGGGHHGHHHFHGHGFGGHHHHHHHHHFGHHGGGEGGGGGGGGGGGCC.<<G0DGGFC:CCC0CC.::00:<00:9C####

Index

@130110_A5T23:1:1101:15646:1320:1#0/2
NTTGTATC
+
#>>AAAFF
@130110_A5T23:1:1101:17282:1320:0#0/2
NTTAGTCT
+
#1>>AAFF

[Devon Ryan]

Have you just tried using FastqMultx from ea-utils (http://code.google.com/p/ea-utils/wiki/FastqMultx)? You'll have to make a file specifying index:sample relationship, but it seems otherwise straightforward and supports paired-end demultiplexing.

[Dan D]

+1 I didn't know about this tool. This looks like the best way to go.

[Dan D]

So all you have available is two FASTQ files, one for the index read and one for the single read, is that correct? Because if you have the basecall data we can run CASAVA. If these two files are all you have I can code up something for you Friday morning.

[peris]

Hi Deedee, Thanks for you reply. Yah you are right, I have only the fastq files, not the raw files. And this is actually a paired end run. I have Read1, Read2 and Index ;these three fastq files. It will be of great help if you can write a code for me as I dont know to write codes for such conditions. Thanks and regards

Answer 1

[yinghua]

I recently wrote a small code for our miseq machine to demultiplex I1, R1, R2 reads assuming barcode reads are in the I1 file. It does a little more than exact matching. For barcodes reads not exactly matched with designed barcodes, which normally amount about 10% of total reads, edit distances with designed barcodes are calculated and a cut off is set by both user and the minimum edit distance among designed barcodes.

https://github.com/yhwu/idemp

Answer 2

[Matt Shirley]

I came up with a solution for this problem a while back, although I didn't know about FastqMultx, and might have done something different now