Getting Warnings When Configuring The Amos Pipeline

Question

[cyclinb-cdk1]

Hi, I'm an undergraduate student trying to go a bit above and beyond on her first ever bioinformatics lab report. Basically, I have a set of shotgun reads that I'd really like to assemble using AMOScmp and visualise in Hawkeye. I have installed MUMmer3.23, apparently successfully, but whenever I try to ./configure AMOS, it configures okay but it gives me multiple warnings, which read something like:

---

WARNING! nucmer was not found but is required to run AMOScmp and minimus2 install nucmer if planning on using these programs

WARNING! delta-filter was not found but is required to run AMOScmp-shortReads-alignmentTrimmed install delta-filter if planning on using AMOScmp-shortReads-alignmentTrimmed

WARNING! show-coords was not found but is required to run minimus2 install show-coords if planning on using minimus2

WARNING! BLAT was not found but is required to run minimus2-blat install BLAT if planning on using minimus2-blat

WARNING! Qt4 toolkit was not found but is required to run AMOS GUIs install Qt4 or locate Qt4 with configure to build GUIs see config.log for more information on what went wrong

WARNING! Boost graph toolkit was not found but is required to run parts of the AMOS Scaffolder (Bambus 2) install Boost or locate Boost with configure to build Scaffolder see config.log for more information on what went wrong

WARNING! Statistics::Descriptive Perl module was not found but is required to run some AMOS scripts

---

Is it just that AMOS can't see that I've installed MUMmer for some reason? Does anyone know how to fix this? Thanks so much in advance!

[Vitis]

Most likely the required executables like nucmer etc are not in your PATH so other programs couldn't locate them, try setting up your PATH parameter in the 'bashrc' file and see what happens. The last warning is about perl module location, which could also be set up in 'bashrc'.

[Ron]

Hi,

Were you able to solve this issue?

Answer 1

[SES]

The nice thing about AMOS is that it combines a lot of related software tools in one package, but this also means there are quite a few dependencies. If you are inexperienced with Unix and compiling software, this will likely not be an easy process to get right. These warnings are telling you that MUMmer could not be found, and neither could Qt, a C++ library, and a required Perl module. Take a look at the "Dependencies" section of the README file included with AMOS. Note that you don't need to install all of these dependencies to use only one of the tools, but it's probably good to try and get a working installation to save yourself some debugging in the future. If you think you installed some of these dependencies, it could be that they are in a custom location. If you type ./configure --help in the AMOS directory, it will tell you how to specify the PATH to Boost, Qt, etc. on your machine and set the environment variables for the programs needed by AMOScmp (i.e., nucmer). In addition, you will probably have to edit one or more source files to get it to compile (at least, that was my experience on Ubuntu and Fedora in the past).

Honestly, I think this may be a bit ambitious for a lab report if you are not experienced in this area (especially if you have time constraints), so don't feel discouraged in any way by considering alternatives. You can find many alternatives for assembly programs and assembly viewers by searching the forum (or google). I encourage you to continue with AMOS because it's great, but my advice would be to also consider other assemblers and maybe use something like Tablet (a very nice program) for viewing the assembly.

[cyclinb-cdk1]

Hi, thanks so much for your help! It is a bit ambitious for a lab report, but I'm going to be doing a systems biology project this summer so I decided it might be good to try and get some familiarity with these things. As interested as I am in it, though, I probably don't have unlimited time -- exams are right around the corner.

I mostly decided to go with AMOScmp originally because I wanted a comparative assembler and it's pretty universally the top hit on Google. Are there other comparative assemblers that are simpler to install/use? I used CAP3 originally and generated the assembly fine, but I wanted to run the same data through a comparative assembler and compare the two using the N50 statistic.

[SES]

So, I take it you have a reference? If so, you are correct that AMOScmp is far more appropriate for this task than CAP3, but the species you are working on is a really important for consideration of the correct program also. If it is a bug of some sort, I think MIRA can do a reference-based assembly but I'll have to think about the others....perhaps someone else will respond on that point.

[cyclinb-cdk1]

It is a bug of some sort -- Staph aureus subspc. aureus, strain USA300_FPR3757 to be precise. There is a well characterised reference sequence for the USA300 strain and several other highly related strains.

Because the purpose of the first part of the lab was to identify the strain, they had us use CAP3 because it was fairly straightforward, could be run in a large-group lab setting, and gave a good enough result that a BLAST of the assembled sequence would identify USA300_FPR3757. However, we're basing our lab (and are using the genome data from) an actual Cambridge, UK outbreak (detailed here: http://www.ncbi.nlm.nih.gov/pubmed/23158674). Because whole-genome sequencing was used alongside traditional epidemiological and antibiogram profiles to understand the dynamics of that outbreak, the scientists knew they were dealing with a Staph aureus strain. Thus, they used a comparative assembler in their study, and it seems to me that is the far better choice.

I will look into MIRA -- could SMALT (Sanger) also be effective in this case?