150bp-1.3kb insert size PE on HiSeq

jmugford · 04-18-2012, 06:36 PM

I have a library that I am running on a HiSeq2000 for PE-100 reads. The fragment sizes range from 150bp to 1.3kb. I can't do much about this range since we are PCR capturing unknown regions using inverse PCR. The majority of the fragments sit in the 500bp range, but will having this wide of a range affect clustering/sequence quality/number of reads? Will there be a bais away from the larger fragments?

Thanks.

HESmith · 04-19-2012, 07:07 AM

Yes, there will be a bias against the larger fragments and a complete absence above a certain size (perhaps 800bp? check with Illumina). Unless the ends of your amplicons need to be matched, you could size-select, shear, and construct libraries from the larger fragments (analogous to sample prep from chromatin IPs).

pmiguel · 04-19-2012, 07:19 AM

Our FAS mentioned off-hand that she had made and sequenced 1.5 kb Illumina PE libraries. Anyone else doing that?

Our first mate end library made for an Illumina sequencer gave much shorter pair distances than expected. 1.4 kb instead of the expect 7 kb. (Needed to do a size selection prior to circularization, I suspect.) Nevertheless adding these reads to normal 400 bp PE reads dramatically improved a de novo fungal genome assembly. If we could get 1.5 kb PE reads -- that would be a game changer for us. I had just assumed it was impossible, though.

--
Phillip

HESmith · 04-19-2012, 09:01 AM

We've sequenced a few libraries w/ 600bp inserts, and all of the metrics (signal intensities, %PF, S/N, Q scores, error rates, etc) were significantly worse than standard 200bp libraries. Also, we clustered at lower densities (~400K) per Illumina's recommendation. We used v3 kits but an earlier version of RTA (whatever was current 6-8 months ago); perhaps the latest update is better suited to larger libraries.

Jon_Keats · 04-19-2012, 10:44 AM

It generally works as but clustering is even more erratic that typical and max densities of 600k are the best you might see. I would worry about such a mixed population and the larger fragments being out competed.

jmugford · 04-19-2012, 08:43 PM

Great - all responses have been helpful. Thanks.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Insert size != Fragment size?	Boel	Bioinformatics	6	12-12-2013 09:28 AM
About Insert, Insert size and MIRA mates.file	aarthi.talla	454 Pyrosequencing	1	08-01-2011 02:37 PM
insert size	polystone	Sample Prep / Library Generation	0	05-04-2010 11:07 AM
Double AMPure size selection in Titanium 3kb span paired end library	H5N1	Sample Prep / Library Generation	5	05-03-2010 11:34 AM
insert size	adrian	Bioinformatics	1	03-18-2010 05:55 PM

04-18-2012, 06:36 PM	#1
jmugford Junior Member Location: Chapel Hill, NC Join Date: Feb 2011 Posts: 4	150bp-1.3kb insert size PE on HiSeq I have a library that I am running on a HiSeq2000 for PE-100 reads. The fragment sizes range from 150bp to 1.3kb. I can't do much about this range since we are PCR capturing unknown regions using inverse PCR. The majority of the fragments sit in the 500bp range, but will having this wide of a range affect clustering/sequence quality/number of reads? Will there be a bais away from the larger fragments? Thanks.

04-19-2012, 07:07 AM	#2
HESmith Senior Member Location: Bethesda MD Join Date: Oct 2009 Posts: 509	Yes, there will be a bias against the larger fragments and a complete absence above a certain size (perhaps 800bp? check with Illumina). Unless the ends of your amplicons need to be matched, you could size-select, shear, and construct libraries from the larger fragments (analogous to sample prep from chromatin IPs).

04-19-2012, 07:19 AM	#3
pmiguel Senior Member Location: Purdue University, West Lafayette, Indiana Join Date: Aug 2008 Posts: 2,317	Our FAS mentioned off-hand that she had made and sequenced 1.5 kb Illumina PE libraries. Anyone else doing that? Our first mate end library made for an Illumina sequencer gave much shorter pair distances than expected. 1.4 kb instead of the expect 7 kb. (Needed to do a size selection prior to circularization, I suspect.) Nevertheless adding these reads to normal 400 bp PE reads dramatically improved a de novo fungal genome assembly. If we could get 1.5 kb PE reads -- that would be a game changer for us. I had just assumed it was impossible, though. -- Phillip

04-19-2012, 09:01 AM	#4
HESmith Senior Member Location: Bethesda MD Join Date: Oct 2009 Posts: 509	We've sequenced a few libraries w/ 600bp inserts, and all of the metrics (signal intensities, %PF, S/N, Q scores, error rates, etc) were significantly worse than standard 200bp libraries. Also, we clustered at lower densities (~400K) per Illumina's recommendation. We used v3 kits but an earlier version of RTA (whatever was current 6-8 months ago); perhaps the latest update is better suited to larger libraries.

04-19-2012, 10:44 AM	#5
Jon_Keats Senior Member Location: Phoenix, AZ Join Date: Mar 2010 Posts: 279	It generally works as but clustering is even more erratic that typical and max densities of 600k are the best you might see. I would worry about such a mixed population and the larger fragments being out competed.

04-19-2012, 08:43 PM	#6
jmugford Junior Member Location: Chapel Hill, NC Join Date: Feb 2011 Posts: 4	Great - all responses have been helpful. Thanks.