Hi Hamid,

I think the maximum amount of differential barcoding in Illumina systems is 96. They offer a plate for that purpose (https://www.neb.com/products/e6609-nebnext-multiplex-oligos-for-illumina-96-index-primers#Product%20Information ).

Depending on the amount of reads per library you want to achieve, you will have to go for multiple sequencing rounds anyway, so you could potentially use these plates and go for 5 -6 library preps. You would have to be aware of potentially confounding effects of this procedure, however, that's the best I have in mind.

Good luck!

Hey Lukas,

Thank you for your response but I know about these kits of 96 , they even offer another method for 4 times 96 indexes, but I am interested in a bigger number of indexes, let say 500 to 1000 or even more. Thats why I am asking.

Thank you anyway,

Hamid.

Alright. However, keep in mind that you want to achieve a certain coverage with your experiment. HiSeq generates around 190 million reads, as a recommendable default amount of reads per library I would go for about 10 million reads. This means that you can only go for 19 libraries in one sequencing run. Using 500 indexed libraries, you end up with 380.000 reads per library. It is doubtful that this is sufficient for any application. So why do you want so many indexes? You (probably) don't want to pool them anyway.

Hi Lukas,

Thank you again for your response. 380k reads for my experiment is way too much... If I obtain between 3-5 k reads, that will be the optimum number for me, that is why I need to pool 500, 1000 or even more samples....

Thanks again

Yes, it is possible. It just depends on the number of barcodes you use. We have 96x96 barcodes in operation and routinely use half of them on one HiSeq flow cell (4608 samples). I am not aware of a company selling more than 384 barcodes but I have not searched for it. We described the sequences of the primers needed for barcoding here: Article Cost-efficient high-throughput HLA typing by MiSeq amplicon sequencing

The demultiplexing server (fastq generation) needs to be set up to allow very many open filehandles. Otherwise it should work without problems.