Extracting reads from barcoded pooled samples

[megh]

Hi,

I'll be glad to receive any comments and suggestions regarding the problem described below. Recently we have completed a small scale sequencing project using 454 FLX platform. Altogether we have 100 samples and they were pooled in row-column format in a way that each sample is present in at least two pools with unique neighboring samples. The objective was to use less number of barcodes to minimize cost and lab work. We formed 10 row pools and 10 column pools. For example: (C = column pool, R = row pool)

C1 C2 C3 C4 C10
R1 1 2 3 4 ..... 10
R2 11 12 13 14 ..... 20
R3 21 22 23 24 ..... 25
....
....
R10 91 92 93 94 ..... 100

So, R1 = 1,2,3,4.....,10
C1= 1,11,21,...,91 etc.

We have used 10 barcodes for both column and row pools. Each sample is present in two pools, for example sample 1 is present in C1 and R1. The sequencing run was with fair quality 454 reads. Our objective is to extract the reads specific to each sample and assemble it. Now we find it difficult to de-convolute. In the ideal case we expected when assemble the reads from pools R1 and C1 and match the contigs against each other we would get the contig/s specific to the common sample (sample 1 in this case) between them.

Any thoughts how we can trace the reads specific to a particular sample and assemble it in the above mentioned scenario?

Thanks.

[vaibhav]

try some 454 sequencing analysis tool

Hi megh,
We also have the similar data, and in our case we had the unique sequence attached to each sample through which we could easily separate from the pool using a small perl script.
Another way of doing it is using 454 analysis tool. We used Roche analysis softwares and there was option in the software to separate the samples from the pools.
Try and see if it can work you also
All the best ..

Thanks and regards

Vaibhav Jain
Research scholar
IGIB, Delhi, India