LastGen Assembly

[dan]

Hi, apfejes

Thanks for your reply (below). I thought I would shift this discussion from the 'Introductions' section to a new post here, seeing as it gets into some specifics...

Hi Dan, welcome to SEQanswers.

I'll take a quick shot at answering your first question, though I should mention that this forum, in general, is for people using the Next-Gen platforms, so this might be the wrong place to be asking these questions. (I've never really done any serious capillary based sequencing or assembly.) Regardless, we're all friendly here, I believe, and I'm certain that some of the members migrated to the next-gen after years on the "Big Dye chemistry" platforms. Hopefully one of them will chime in and give you better answers than I can.

OK, its not that bad, I do have the broad idea, but where can I find out more? What books should I be reading? Which websites have the best information? How does this kind of sequencing compare to NextGen sequencing in terms of speed, throughput, cost, coverage, de-novo 'assemble-ability', etc?

If you're asking about non-Next-Gen sequencing, you're basically referring to all of the sequencing done before the next-gen platforms arrived in 2006/2007. If you pick up any reasonable molecular biology or biotech textbook, it'll probably have a few paragraphs on it. (Look up Dideoxy or Sanger DNA sequencing for the chemistry, and capillary sequencing for the machines. I'd be surprised if you couldn't find a few hundred web pages on it - there are nearly a million on capillary sequencing.

In terms of cost, Sanger sequencing can be an order of magnitude (or more) expensive per base, but has some very good features: It's accurate, it's targeted (using primer pairs) and is a trusted method. The key is that it's not competing with Next-Gen sequencing - they have very different applications. Now that Next-Gen is available, I don't think it's particularly cost effective to sequence a genome using Sanger sequencing, but it has been done (eg. the human genome), although pretty much everyone doing next gen work will use Sanger sequencing to verify any predictions they make.

Generally, I would sum it up as this: Sanger sequencing is used to look at a single site of dna, (eg, a BAC) with great specificity and for reads of about 1000bp in length. Next-Gen sequencing is more of the "pick X million random locations" type (length and number of sequences depend on the technology used), which wouldn't make sense if you wanted to look at a single BAC.

(Or course, if you have access to next-gen sequencing, you wouldn't be making a BAC library in the first place.)

As for information, I suspect another good place to start is pubmed. Papers before 2006 will all be discussing how they did assembly with this type of sequencing information. I'm certain there are many applications out there to assist in this task. Their manuals would also be full of helpful hints.

Hopefully that's enough to get you pointed in the right direction, though I've left most of your questions unanswered.

Good luck

No problem! Its great to get the information. I think the existing BAC library (including BAC end sequences) [REF] in addition to an 'Ultra high density genetic map' [REF] in potato has kept the consortium on the path of 'LastGen', however, there is the possibility of combining approaches to help finish the project.

Coincidentally(?) I got an email from Bioinformatics World Wide with the following two very useful publications:

• Algorithmic challenges in mammalian whole-genome assembly
• Errors in sequence assembly and corrections

Within these I found the following useful citations

• Genomic mapping by fingerprinting random clones: a mathematical analysis. - Looks key.
• A whole-genome assembly of Drosophila - ;-)
• Automated DNA sequencing of the human HPRT locus.
• Assembly of the working draft of the human genome with GigAssembler.

I thought I would post the above back to the forum in case people find themselves in a similar situation (i.e. having to understand LastGen data in a NextGen community :-)

Thanks again for your help,

Dan.