1
$\begingroup$

I need to align bacterial genomes (namely mycobacteria, over 4 Mb in length), and my choice would have been either 'muscle' or 'clustalW'; however these algorithms are too memory demanding for such long sequences thus I can't run the multiple alignment on my machine.

Is there any other program that I could use to align multiple bacterial sequences?

Improve this question
$\endgroup$
2
  • $\begingroup$ There are many others. I'm surprised you run out of memory... How much RAM do you have? If working on Linux, swap space is easy to allocate and can be used as a sort of memory if time isn't a constraint. Since this is the issue, please do provide information on what biology you are looking for with the sequence alignment, and what are the exact memory limitations. What specific errors do you get. Be aware that this site aims to address biological problems, not computational ones, so do shift the emphasis onto maintaining biological accuracy rather than overcoming memory constraints. $\endgroup$
    – James
    Commented Nov 1, 2016 at 9:53
  • $\begingroup$ Ok, I'll try to keep it on the biology rather than bioinformatics; however I have 15 Gb of RAM. $\endgroup$
    – Gigiux
    Commented Nov 3, 2016 at 22:48

2 Answers 2

Reset to default
5
$\begingroup$

First of all, do not, I repeat not, use ClustalW. That is horribly out of date and was never very good to begin with1. It was just the first to become common. At least use ClustalΩ. Anyway, while I haven't done much aligning in the past few years, back when I did, I got the best speed/quality ratio from MAFFT.

MAFFT has been used in various sequencing projects and can deal with whole genomes easily. Whether or not it will run correctly on your machine depends on the specs of your machine, of course. However, it should be considerably more efficient than clustalW, although not necessarily more than muscle.

Before doing any of this though, are you sure you need to align whole genomes? Could this be an XY problem? You might want to consider posting a new question here and explaining what your final objective is. Also, if these are published genomes, you might be able to find an existing alignment.

If you really do need to run a whole genome alignment and your machine can't use MAFFT or muscle or any of the other modern aligners, you need a new machine. No matter how efficient your aligning algorithm, a genome is a lot of data and you need a strong machine with serious RAM for it. If you are on a *nix machine, you could ask about how to create a swapfile to increase your memory on Unix & Linux.

1Don't get me wrong, it was state of the art and everything back in the day, but the state of the art wasn't very good, that's all.

Improve this answer
$\endgroup$
3
  • $\begingroup$ The issue of aligning long sequences came out already somewhere else (actually from the developer of muscle); what is such issue? If I have related species (in this case the members of the Mtb complex) doesn't it make sense to align the whole genome and see where differences in genome (especially indels) arise? Isn't this the reason of being of the Whole Genome Sequencing research? $\endgroup$
    – Gigiux
    Commented Nov 3, 2016 at 23:21
  • $\begingroup$ @Gigiux sorry, for some reason I never saw this comment. The alignment you need depends on the question you want to answer. Yes, sometimes whole genome alignments are useful. Often, however, they are not and you are only interested in specific areas. I can't help you choose unless you explain exactly what you are trying to do. $\endgroup$
    – terdon
    Commented Nov 29, 2016 at 13:22
  • $\begingroup$ I think I understood why whole alignment is not good as such for inter-species comparison: regions of homology can be dispersed in different parts of the genomes thus there is no co-linearity between sequences. I have found Mauve an excellent tool for inter-species alignments. $\endgroup$
    – Gigiux
    Commented Jan 21, 2017 at 9:18
0
$\begingroup$

I have used LASTZ in the past, you can set the word size and thereby control memory usage. You might be better off asking your question on Biostars, there are more skilled bioinformaticians lurking around there.

Improve this answer
$\endgroup$
1
  • $\begingroup$ I'd say there are skilled bioinformaticians here too so as not to offend this community ;) but you're right. Here we're more interested in the underlying biology than at biostars, which is a biology SO mashup. $\endgroup$
    – James
    Commented Nov 2, 2016 at 7:37

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .