Today a colleague of mine asked the following question:

" Assuming I need to build from 0, a chromosome of a fish, with short reads but no other reference whatsoever [de novo assembly]:

  • how much work is that?
  • Is there a generic software (like SAMtools) that will align the reads in a scaffold one can use?
  • Basically, given a reasonably clear pipeline in terms of software, is it still blood sweat and tears or is it just a matter of getting it on a cluster?"

Very grateful for any suggestions, sources of information, software etc.

  • 1
    $\begingroup$ tl;dr: It’s blood, sweat and tears. I haven’t been involved in this myself yet but I know people who are. Depending on the complexity of the genome, you’ll get tons (they got in the order of 1M) non-overlapping contigs. $\endgroup$ – Konrad Rudolph Dec 7 '13 at 19:34
  • $\begingroup$ would it really be necessary to use 30bp reads? that's very short. $\endgroup$ – Thomas Ingalls Dec 9 '13 at 16:20
  • $\begingroup$ cheers guys, actually I pressumed ~30bp to give an idea but perhaps it is more flexible what lengths do you have in mind @ThomasIngalls? Hmm I wonder what the blood sweat and tears would be caused by? I presumed there would be some standard pipelines to automat the whole shebang? $\endgroup$ – hello_there_andy Dec 9 '13 at 23:18
  • 1
    $\begingroup$ De novo sequencing without using some long-read techniques for cleanup (like Sanger sequencing) requires a lot of coverage, and you still end up with substantial uncertainty in the results, due to the large amount of repetitive sequences that are longer than 30bp in most genomes. The reply by @bitwise gives a great idea of what I was thinking. $\endgroup$ – Thomas Ingalls Dec 15 '13 at 19:13

If you only want to use only sequencing techniques, you have a problem.

To get a feeling of what kind of results to expect, consider this paper published recently in Nature Genetics. They tried to assemble a whale genome de novo. They had 7 (!) paired-end libraries with different insert lengths ranging from 170bp to 20kb. Read lengths were mostly 100bp and in some cases 49bp. Average genome coverage was 91x.

Assembling this extensive data, they end up with over 100,000 contigs when the assembly is done.

So you really can't get a high-quality complex (i.e. large) genome assembled from only short-read sequencing data using the standard techniques.

That said, recent approaches such as libraries with much longer reads lengths (here) or the use of Hi-C data (here and here) do offer a way of getting high-quality complex genome assemblies using only sequencing data.

  • 1
    $\begingroup$ "If you only want to use only sequencing techniques, you have a problem." -- Do you mean resequencing techniques, or high-throughput sequencing techniques? $\endgroup$ – Thomas Ingalls Dec 15 '13 at 19:13
  • 1
    $\begingroup$ @ThomasIngalls I mean assembling a complex genome de novo using high-throughput sequencing. Resequencing assembly is not de novo assembly. $\endgroup$ – Bitwise Dec 16 '13 at 9:21

You can try looking around biostars.org, which is like stackexchange, but for bioinformatics.

Velvet is one example of a de novo assembler.

But 30 bp is really short, and animals have big genomes (not as tough as lots of plants and fungi, but still tough)

What you would get is a bazillion short contigs. It would not be pretty.

  • $\begingroup$ cheers for the Velvet recommendation $\endgroup$ – hello_there_andy May 14 '18 at 12:32

I really like the genious software suite. It can multithread and really use the performance of your computer. Even complicated things like De Novo assembly are very very intuitive.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.