>100bp hiseq read length ?

[dvh]

Hi,

Has anyone experience of >100bp reads with HiSeq1000/2000 ?
(TruSeq v3 chemistry)

We could make good use of slightly longer reads than 100bp.
We are also going to waste reagents from a 200 cycle kit otherwise (only want single end read + index).

Wondered if anyone had any data. If so what the qualities are like at 3' end of reads etc.
I do realise Illumina only officially support 100bp.

regards

david

[rflrob]

I know the sequencing center here at Berkeley has the option for 150bp, both single and paired end (although looking back through the calendar I don't know if anyone's done 150 single end). You might try getting in touch with them for advice.

[kmcarr]

We are also going to waste reagents from a 200 cycle kit otherwise (only want single end read + index).

david

When we run an SR 100bp + index we mix 2 50 cycle kits (each good for at least 57 cycles) so there's little waste.

[pmiguel]

Hi,

Has anyone experience of >100bp reads with HiSeq1000/2000 ?
(TruSeq v3 chemistry)

We could make good use of slightly longer reads than 100bp.
We are also going to waste reagents from a 200 cycle kit otherwise (only want single end read + index).

Wondered if anyone had any data. If so what the qualities are like at 3' end of reads etc.
I do realise Illumina only officially support 100bp.

regards

david

If you are going to waste the reagents anyway, go for the 150 nt read. Why not? If you do, please post your results here.

BTW, I would bet that most places offering 150 nt Illumina reads are doing those reads on a GA-IIx. (Or a MiSeq...)

--
Phillip

[peromhc]

The 150bp pe is crappy-- here is a plot from 1/5 (40 million reads) of a recent v3 HiSeq run. Note the 3' end of the read is >>5% error.

For me, the 100PE is the way to go, after getting these results.

[pmiguel]

The 150bp pe is crappy-- here is a plot from 1/5 (40 million reads) of a recent v3 HiSeq run. Note the 3' end of the read is >>5% error.

For me, the 100PE is the way to go, after getting these results.

Maybe out to 125 nt? At lower densities (less than 500 K/mm2) quality might stay above an acceptable level out to 150?

I tend to agree with you though. Not sure there is much utility in scraping another 50 nt from a run, when you could be 1/2 through a new run by the time it finished.

--
Phillip

[dvh]

thanks for your graph peromhc.

it does look like we might be able to make use of HiSeq+Truseq v3 data out to 125-130bp as pmiguel suggests.

for the application we want (amplicon sequencing with bidirectional sequencing primers) longer reads offer something more reads (or paired reads) do not !

regards, dvh.



(ps: we tried GAIIx 150bp reads, with multiple lanes loaded with 1 to 16pM of the same library. Cluster density seemed to make no difference to 3' base call quality, surprisingly. Not currently the same TruSeq chemistry as HiSeq though)