Scientific

Deep Sequencing Notes


Compiling the Solexa Pipeline v 0.2.2.6

The solexa pipeline converts images to fragment sequences. Beware that the space requirements are probabhly larger than you expect:

Each lane is measured individually and placed in a folder L00<lane>
In each folder we have the different cycles called C<cycle>.1
In each cycle we have around 330 scanned tiles listed as s_<lane><tile><base>.tif
Each image is 2Mb, so in total we have

2Mb * 4 (bases) * 330 (tiles) * 8 (lanes) * 36 (cycles) = 760320 Mb = 742 Gb
per experiment.

The paired end module might potentially double this to 1.45 Tb per experiment.

On Linux

Go to the top of the directory

On a macbook pro

FFTW seems to be at version 2.1.5 on the standard install and that does not work.
Download fftw 3.1.2.tar.gz from
http://www.fftw.org/download.html
Unyip it and go to the direcrtory fftw-3.1.2
Enter ./configure --enable-float
./make
sudo make install

which will install the fftw3 package in /usr/local/lib/

Add the line
LDFLAGS += -L/usr/local/lib/
at the top of
/Users/wernerv/Solexa/Pipeline/SolexaPipeline-0.2.2.6/Firecrest/Makefile

Solexa Pipeline v0.3

Mac

Seems too complicated to get it to work. Most of the packages are too old.

Linux

apt-get install gnuplot
apt-get install libbz2-dev  (the bzlib library and headers)
apt-get install libxml-simple-perl (this is the XML::Simple perl module)
apt-get install imagemagick (provides convert)

To then compile the software you should remove the dependecy files and object files in the firecrest/Matrix directory. Practically this is done with rm *.d *.o

Don't forget to type make install, otherwise the goat directory will be missing a lot of useful executables.

Also not that the goat folder should be pointing to some of the proper binaries and those links need to be made manually:

cd Goat
ln -s ../Firecrest/run_image_analysis .
ln -s ../Firecrest/analyse_image .
ln -s ../Bustard/run_base_caller .
cd Matrix
ln -s ../../Firecrest/Matrix/savinio .
cd ../Phasing
ln -s ../../Bustard/Phasing/phasing_estimate .

More Deep Sequencing notes
- http://analysis.yellowcouch.org/