Structural Variation Detection Software Using Next Generation Sequencing

Question

[Junfeng]

Hi, I would like to detect structural variation using next generation sequencing. Now there are many methods are available (refer to these review papers:

• (1) Medvedev P., Stanciu M.,Brudno M. Computational methods for discovering structural variation with next-generation sequencing. Nat. Methods 2009;6:S13-S20.
• (2) Ruibin Xi, Tae-Min Kim and Peter J. Park. Detecting structural variations in the human genome using next generation sequencing. Briefings in Functional Genomics; 2010; 9(5-6): 405-415.

I would like to know what is the suitable structural variation detection software for my current job.

I have tried some methods, such as SVDetect (http://svdetect.sourceforge.net/Site/Home.html) and BreakDancer (http://sourceforge.net/projects/breakdancer/). But I have problems when I run these programs. For example, when I run breakdancer_max, it displays that

 "./breakdancer_max: /usr/lib64/libz.so.1: no version information available (required by ./breakdancer_max)

./breakdancer_max: /usr/lib64/libstdc++.so.6: version `GLIBCXX_3.4.9' not found (required by ./breakdancer_max)"

These problems cannot be solved by our server administrator.

When I run SVDetect, I found that the program becomes very slow after it runs several hours (since the program occupy too much memory). Do anyone know how to improve the computational speed of this program or tell me what is the suitable program for my project? Thank you.

Answer 1

[Drio]

1. The binary version that you downloaded is expecting certain libraries that are not available in your system. You want to install those libraries in your OS. Look for zlib and libstdc++. Notice that you may have those libraries already installed but with newer versions. In that case you want to download the source version of the tool and compile the binary yourself. Your sysadmins can help you on that.

2. Read carefully the basic algorithm used in the tool. If it uses a lot of memory (it shouldn't) and the OS starts swapping in/out (meaning it has to start saving pieces of memory to disk) the performance will degrade dramatically. The only solution would be to compute the SVs in another machine with more memory.

[Junfeng]

Thank you very much for your kind reply. It is very useful!