README 19 June 2011

Spanner is a c++ program for the detection of Structural Variation events from whole genome sequenced read pair data. It uses aligned read information in bam format and was developed in conjunction with the Mosaik alignment program (https://github.com/wanpinglee/MOSAIK).

Compiling the source code

git clone [email protected]:chipstewart/Spanner.git (this repository) to an appropriately named subdirectory (eg ~/Spanner) on a linux or macOSX computer.

Spanner links to two other libraries:

1. Bamtools

If bamtools is not already installed Install bamtools on your computer. Bamtools and the corresponding installation instructions can be found at: git clone https://github.com/pezmaster31/bamtools.git

2. RE2 regular expression library

RE2 and installation readme instructions can be found at: hg clone https://re2.googlecode.com/hg re2

In addition to Bamtools and RE2, Spanner also uses the TCLAP command line parser. TCLAP is not a library, but rather a set of include header files. The TCLAP header tree should be installed on your include header path (eg. /usr/local/include). If this is not possible then TCLAP should be included as a qualifier on the CPPFLAGS line of the Makefile:

CPPFLAGS=-Wall -O2 -march=nocona -I$(RE2_ROOT) -I$(BAMTOOLS_ROOT)/include -I$(/home//include)

INSTALLING SPANNER

Go the area with the Spanner source and find the Makefile. The Makefile needs to know where to find Bamtools and RE2, so it is necessary to edit the Makefile to change RE2_ROOT=" " and BAMTOOLS_ROOT=" " to point to the main install areas for those packages. Once RE2_ROOT and BAMTOOLS_ROOT are set, then type "make". I hate to predict what will happen, but if we are lucky a Spanner binary will appear in the directory despite gcc complaints about deprecated header files.

Run Spanner without arguments. You should see something very much like:

begin Spanner!      Version 7.00
command line:
Spanner
hey ... need some input file(s) !
 try  -1 and -2 , or -i

If you don't see this, then we'll need to fix the compile & link on your server. I hope we can avoid that because I use a couple of libraries that would need to be installed (bamtools and RE2).

Even more informative would be to run Spanner -h which will produce a list of all the command line arguments that Spanner can accept. Spanner has years of historical developement behind it so some options are more obscure than others. The command line options are listed by Spanner -h in no particular order.

Spanner processing steps:

1: Build spanner data structures (*.span files) and split by chromosome.

1a. A short step between 2 and 3 in which multiple sequencing runs/lanes are combined. Spanner combines runs by loading all build files from a subdirectory. A python script (SpanLinks.py) can make linux symbolic links in a subdirectory from all build files from all runs/lanes that should be combined.

2: Detect SV's: detect clusters of fragments spanning SV breakpoints and write the corresponding event list of SV's. The event list is a VCF-format text file with many fields in addition to event type, chromosome, position, and length. Depending on the SV type, there are up to 20 other descriptive fields in the output event lists for each event. The other fields are parameters that I sometimes look at but are not generally relevant.

At the moment, the output SV list includes deletions, tandem duplications, inversions, cross-chromosome translocations, and mobile element insertions detected by read-paired (RP) constraint.