This software was written by David Quigley in the Balmain Lab at UCSF.


**********************
BUILDING BOOST 
**********************

Boost source code is available at boost.org

ON OSX
--------------------------------

IMPORTANT FOR USERS OF OSX 10.9: Build with libc++ explicitly, and build the 64-bit 
version. To build boost on OSX 10.9, I have used:

sudo ./bootstrap.sh --with-libraries=system,filesystem,regex,thread
sudo ./b2 link=static runtime-link=static toolset=clang \
      cxxflags="-stdlib=libc++" linkflags="-stdlib=libc++"

ON WINDOWS
--------------------------------

It may not be necessary to build boost for Windows, as once again the binary files can be 
downloaded from boost.org. I am using Visual Studio 2005 (VC8), and because of compiler 
incompatibilities with the pre-built binaries I found it necessary to build a 32-bit 
version of boost on Windows. In order to call bootstrap.bat, For reasons I do not 
understand I had to define two symbols in builtins.c that should have been present 
somewhere in windows.h:

#ifndef FSCTL_GET_REPARSE_POINT
#define FSCTL_GET_REPARSE_POINT 0x000900a8
#endif
# ifndef IO_REPARSE_TAG_SYMLINK
# define IO_REPARSE_TAG_SYMLINK (0xA000000CL)
# endif

After calling bootstrap.bat my call to bjam to build a static, multi-threaded 32-bit 
version of boost was:

bjam link=static address-model=32 threading=multi runtime-link=static -a
     --with-regex --with-system --with-filesystem --with-thread --with-date_time install 

You want in this call to link to the mt-s versions of the libraries:
libboost_thread-vc80-mt-s-1_56.lib 
libboost_system-vc80-mt-s-1_56.lib 
libboost_regex-vc80-mt-s-1_56.lib 
libcarmen.lib

Note that your version of boost (and your build environment) may be different.

I pass the pre-processor definitions 
    WIN32
    NDEBUG
    _CONSOLE

If you're building this way be sure to specify the /MT runtime library flag, not /MD.


*************************
* BUILDING JASPER: OS X *
*************************

Please see the notes in the jasper README.
 
    
***********************************
* The CARMEN software data format *
***********************************

The CARMEN software suite expects data to be formatted into three tab-delimited 
text files: 

1) an expression file
2) a probe attributes file
3) a sample attributes file

It is STRONGLY recommended that neither attribute names nor sample names
contain a space. It is recommended that attribute values do not contain a 
space, simply to make your life easier.

-----------------------------------
The format of the expression file:
-----------------------------------

Column one:
The first row is the word IDENTIFIER.
Subsequent rows contain one unique identifier (e.g. a microarray probeset identifier). 

Columns two and beyond:
The first row is the name of a sample. Each sample name must be unique.
Subsequent rows contain the values that identifier-sample pair.
Missing values are allowed. Missing values should be coded with NA.

Example:

IDENTIFIER	RU109_1000_tail	RU109_1001_tail	RU109_1002_tail
10344624	11.106	10.989	10.748
10344633	9.858	NA	9.467
10344637	10.453	10.332	10.378

----------------------------------------
The format of the probe attributes file:
----------------------------------------

Column one:
The first row is the word IDENTIFIER.
Subsequent rows contain one unique identifier (e.g. a microarray probeset identifier). 
These probe identifiers should match the identifiers in the expression file.

Columns two and beyond:
The first row is the name of an attribute. Each attribute name must be unique.
Example attributes: Chromosome, transcription.start.location, is.refseq
Subsequent rows in each column contain the values for a given identifier-sample pair.
Missing values are allowed. Missing values should be coded with NA.

Example:

IDENTIFIER	Chromosome	transcription.start	strand	symbol
10344624	chr1	4807893	+	Lypla1
10344633	chr1	4858328	+	Tcea1
10344637	chr1	5083173	+	Atp6v1h

------------------------------------------
The format of the sample attributes file:
------------------------------------------

Column one:
The first row is the word IDENTIFIER.
Subsequent rows contain one unique sample.
These sample identifiers should match the identifiers in the expression file.

Columns two and beyond:
The first row is the name of an attribute. Each attribute name must be unique.
Example attributes: p53.mutant, tissue.type
Subsequent rows in each column contain the values for a given identifier-sample pair.
Missing values are allowed. Missing values should be coded with NA.

Example:

IDENTIFIER	scan.date	sex
RU109_1000_tail	june.2012	F
RU109_1001_tail	june.2012	F
RU109_1002_tail	june.2012	M