[marcaill]

Someting like that ?

http://www.pnas.org/content/110/14/5552.abstract

However, very close to the LFR technology...

[KevinLam]

they have 3000 8.5 kb reads now .. and it's avail as a public dataset more details here

http://kevin-gattaca.blogspot.sg/2013/07/what-would-you-do-with-3k-of-8500-bp.html

Done a FASTQC on both sets of data link as above

There's also a diagram explaining the method in the First publication using the Long Read Seq (LRseq) The genome sequence of the colonial chordate, Botryllus schlosseri | eLife Contains a diagram explaining the LRSeq protocol. This experiment yielded ~1000 6.3kb fragments

http://elife.elifesciences.org/content/2/e00569

[Cofactor Genomics]

Patent information on it can be found at:

https://docs.google.com/viewer?url=patentimages.storage.googleapis.com/pdfs/US20130079231.pdf

It reminds me of human genome project clone sequencing at the massively parallel level (with a few tricks to tag the molecules).

Jarret Glasscock
Cofactor Genomics

http://www.cofactorgenomics.com

[ychang]

Does anyone try the method described in the PNAS paper?

http://www.pnas.org/content/110/14/5552.abstract

I am trying it now and want to share the experience with our colleagues!

[ymc]

Can Moleculo help diagnosing triplet repeat disorders like Huntington's Disease?

[ymc]

http://www.ncbi.nlm.nih.gov/pubmed/25188499

This paper used 224Gbp worth of 2x100 reads to generate 4.2Gbp synthetic long reads.

Is 1 in 53 about the yield people are getting?

[pmiguel]

Can Moleculo help diagnosing triplet repeat disorders like Huntington's Disease?

Probably not. Or at least no better than normal Illumina sequence does.

--
Phillip

[pmiguel]

http://www.ncbi.nlm.nih.gov/pubmed/25188499

This paper used 224Gbp worth of 2x100 reads to generate 4.2Gbp synthetic long reads.

Is 1 in 53 about the yield people are getting?

Yes.
The Illumina Long Read Synth method may seem complex, but it is actually fairly straight-forward. You just shear DNA to about 10 kb and ligate some adapters on their ends.

Then you aliquot a very small amount of DNA into each well of a 384-well plate. The amount is up to you, but in the Illumina protocol you would end up with about 200-300 of these 10kb fragments per well. You then PCR amplify these fragments so you have many copies of each and then make a different Nextera library (with a different index) from each well. The you just pool and sequence them.

As a result each well has 2-3 megabases of ~10kb fragments in it. This helps out your sequence assembler because instead of having to traverse every false path across all the repetitive DNA in an entire genome, you just have to deal with a tiny fraction of that genome.

The main down-side is just what you note -- to assemble those 2-3megabase/well from Nextera library reads you want to have sufficient read depths. 50x coverage should be good for most purposes. But, ouch, that is a 50x hit to your final read coverage! Granted the reads should be highly accurate ~10kb reads. They might be just the thing for some projects. But for many projects mate-pair libraries will probably give much more bang for the buck...

--
Phillip

[ymc]

Yes.
The Illumina Long Read Synth method may seem complex, but it is actually fairly straight-forward. You just shear DNA to about 10 kb and ligate some adapters on their ends.

Then you aliquot a very small amount of DNA into each well of a 384-well plate. The amount is up to you, but in the Illumina protocol you would end up with about 200-300 of these 10kb fragments per well. You then PCR amplify these fragments so you have many copies of each and then make a different Nextera library (with a different index) from each well. The you just pool and sequence them.

As a result each well has 2-3 megabases of ~10kb fragments in it. This helps out your sequence assembler because instead of having to traverse every false path across all the repetitive DNA in an entire genome, you just have to deal with a tiny fraction of that genome.

The main down-side is just what you note -- to assemble those 2-3megabase/well from Nextera library reads you want to have sufficient read depths. 50x coverage should be good for most purposes. But, ouch, that is a 50x hit to your final read coverage! Granted the reads should be highly accurate ~10kb reads. They might be just the thing for some projects. But for many projects mate-pair libraries will probably give much more bang for the buck...

--
Phillip

Thanks very much for your detailed reply.

Do you know if this Moleculo method can detect trinucleotide repeats (e.g. repeats of CAG)?

[pmiguel]

Thanks very much for your detailed reply.

Do you know if this Moleculo method can detect trinucleotide repeats (e.g. repeats of CAG)?

Like I wrote, it would be no better at that than plain Illumina reads. Meaning that if the stretch of trinucleotide repeats is long enough, neither Illumina paired-end nor Synthetic Long Reads will be able to traverse long CAG repeats.
It would be no better because the Synthetic Long Read derive from normal Illumina reads after assembly.

--
Phillip

[ymc]

Like I wrote, it would be no better at that than plain Illumina reads. Meaning that if the stretch of trinucleotide repeats is long enough, neither Illumina paired-end nor Synthetic Long Reads will be able to traverse long CAG repeats.
It would be no better because the Synthetic Long Read derive from normal Illumina reads after assembly.

--
Phillip

Oh I see. So it is mostly for phasing and assembly as it is mentioned at Illumina site.

I think conceptually it is possible to truly stitch reads together by adding barcodes intelligently. Of course, it might not be feasible physically and/or chemically.

If 10X can do that, then it can be a true killer app.

[nucacidhunter]

I think conceptually it is possible to truly stitch reads together by adding barcodes intelligently. Of course, it might not be feasible physically and/or chemically.

If 10X can do that, then it can be a true killer app.

This is what they are claiming their technology can do. But, as pmiguel has pointed out increasing the number of wells with limited number of long fragments and barcoding them while converting to shorter fragments for Illumina systems will have limitations that only long fragment sequencing such as PacBio or Nanopore (if they release any product) can overcome.

[ymc]

This is what they are claiming their technology can do. But, as pmiguel has pointed out increasing the number of wells with limited number of long fragments and barcoding them while converting to shorter fragments for Illumina systems will have limitations that only long fragment sequencing such as PacBio or Nanopore (if they release any product) can overcome.

10X announced their GemCode platform but they didn't say anything about triplet repeats. So they are just a better moleculo with longer reads and better yield?

[pmiguel]

10X announced their GemCode platform but they didn't say anything about triplet repeats. So they are just a better moleculo with longer reads and better yield?

ymc, why are you assessing technologies designed to traverse/deconvolute large scale repeats (10kb-100kb) when your focus is very short repeats?

How long are these triplet repeats you are interested in? Can't they be traversed using normal (100-300 base) Illumina reads?

--
Phillip

[ymc]

ymc, why are you assessing technologies designed to traverse/deconvolute large scale repeats (10kb-100kb) when your focus is very short repeats?

How long are these triplet repeats you are interested in? Can't they be traversed using normal (100-300 base) Illumina reads?

--
Phillip

I just think a true synthetic long read should be as close to a true long read as possible. Theoretically I think it is possible to do that by adding barcodes intelligently. Of course I don't know if that's feasible physically or chemically.

I am interested in 200+ triplet repeats. So 1000bp reads should solve most of my problems. Hopefully the next MiSeq can do that.