This page records various problems and issues that I encountered, mainly while working at Warwick, in case they are of any help to other people who might encounter similar probems.
Apple macs and OS-X
Eclipse and GCC
Eclipse does not provide explicit support for gcc precompiled headers. This is a relatively simple way of using precompiled headers.
There are many pages that describe Makefile syntax, and none of them provided quite the information that I was looking for. This provides information on the Makefile syntax that I finally ended up using.
Using Eclipse to debug C++ template classes
The eclipse debugging environment does not show the contents of C++ template classes in a user friendly way. This provides details of how the python addin can be installed to solve this problem.
Rebuilding gdb within Cygwin
There are a number of reasons why gdb may need to be rebuilt when being used within cygwin, particularly to include the python support that is needed in order to view the contents of template classes. This is how it is done.
warning: can't find linker symbol for virtual table for problem when debugging
Bioinformatics software issues
GENE-E is a very impressive peice of software for analysing gene expression data, but the documentation is very sparse and there are many things where the GUI does not make it that obvious how something should be done. My additional notes
Building tophat and cufflinks on Windows
No support is provided for buiding these tools on Windows, but it is now comparitively simple using cygwin64 (not worth doing with 32bit cygwin). Instructions here
Building UCSC tools such as samtools and wigToBigWig for Windows
Again, there is no support for building these on windows, but 64 bit tools can be built using cygwin64. Instructions here
Blast does not find all matches
The ncbi blast program is a basic bioinformatics workhorse for finding the degree of match between two protein or DNA sequences. Blastall finds the best matches between a target sequence and a set, potentially very large, of other protein sequences.
Whilst using blastall to find matches between a possible protein sequences and the sequences in the swissport data set (A 150Mbyte fasta file) it was found that release 2.2.18 only found about 25% of the matches found by release 2.2.22. There is nothing in the blast release notes that draws attention to such a major deficiency in the early releases of blast.
Info on installing the current version of PyMol can be found here.
Default options can be set by making a copy of the .pymolrc_example file and calling it .pymolrc
My PhD made extensive use of the CisGeome toolset for manipulating next generation sequenceing data, which I heavily customised. These pages provide more details about using cisGenome and the customisations
Using 64bit Perl
I found that strawberry perl provided the best supported 64 bit Perl, however there was a problem with installing some modules (e.g. Padwalker) where there was a compile error. It turned out that this was because there were two different gcc compilers on the PATH, and it was using the wrong one. The solution appears to be to modify the cpan.bat file so that the strawberry perl compiler is only on the path when using cpan by adding the SET PATH as below
@rem = '--*-Perl-*--
if "%OS%" == "Windows_NT" goto WinNT
perl -x -S "%0" %1 %2 %3 %4 %5 %6 %7 %8 %9
perl -x -S %0 %*
Cricket: My Wordpress theme