crimap documentation (version 2.4)

3. Getting started

  1. Put all files on the diskette provided into a single directory on your computer.

  2. Compile the source code files (i.e. those whose names end in .c), and link them together with your C stdio and math libraries, producing an executable file called crimap. There are slight differences in the conventions used by some C compilers which may make it necessary for you to modify some of the source files before they will compile correctly; usually these will be evident from the error messages generated by the compiler. For example, if your operating system is casesensitive (e.g. UNIX), you will need to rename all files in lower case (DOS converts all file names to upper case). We have declared some stdio and math library functions in the crimap functions where they are used; if your compiler objects to this, you will need to delete the offending declarations. Note to VAX C users: there is a bug in the optimizer of the VAX C compiler, ver. 2.3 (and possibly in other versions), which causes it to compile the source file "eonv.c" incorrectly. When the program is run, this will result in an incorrectly created phase-known data structure in the .dat file (see below). To avoid this, compile eonv.c with the /NOOPTIMIZE option. All other source files should be compiled with the default /OPTIMIZE option. Also see comments on "underflow" under TECHNICAL NOTES.

  3. The command format for running the program is

    crimap {chromosome number} {option}

    Example:

    crimap 7a twopoint

    The option name must be entered in lower case letters. Chromosome "number" (which may consist of any string of digits, possibly followed by letters - for example 7p, or 17nf, or 0a) may be replaced by the name of the parameter file (described below); for example,

    crimap chr7a.par twopoint

    You must provide a .gen file (named in accordance with the conventions described below, and residing in the same directory as CRI-MAP) which contains the raw genotype data, and run the prepare option first in order to create the other files required by the program. All program output (apart from specific information written to one of the four files described in the next section) is displayed using the "printf" function in C. It will thus be displayed on the terminal (unless redirected to a file by means of commands to the operating system) if the program is run interactively, or written to a log file if the program is run in batch mode. (The latter procedure is the most convenient way to make a copy of the output. In UNIX and some other operating systems, one can simply redirect the output to a file.) The prepare and merge options are the only ones requiring interactive input.

  4. As a test run with the chromosome 7 data set, chr7a.gen, provided with the program, use the command

    crimap 7a prepare

    to create a .dat file and a .par file for subsequent use by the option all, with the loci 2 8 9 10. (Specify any two of these as the "ordered loci", and the other two as the "inserted loci".) Use default values for the other parameters. (NOTE: If the program stops prematurely, displaying the message "Your compiler uses a different size for integers; see documentation for changes that will have to be made in the source code", then you will need to make certain minor modifications in the source files and recompile. See "16 bit integers" under TECHNICAL NOTES, below.

    If the program stops prematurely, displaying the message "ERROR: ALLOCATION FAILED IN MORECORE", then the program's default memory allocations exceed what your operating system can provide. Reduce the value of nburlloc in the .par file (using prepare, or a text editor) and see "Memory management" under TECHNICAL NOTES, below.

    Then run CRI-MAP with the command line

    crimap 7a all

    The program prints out parameter values, tests all 12 possible orders of these 4 loci, and finally prints out a sorted list of all orders whose log10 likelihood is at least that of the best order, less 3.0.

    You should get the answer: 

              2 10 9 8   -97.937
          2 10 8 9  -100.839

    The run to evaluate these 12 orders should take less than a minute on most workstations or moderately powerful personal computers (for example, about 50 secs on a Mac II, about 25 secs on a MICROVAX II without a numerical coprocessor, about 8 seconds on a SUN 3/60 or 2 seconds on a 4/60; exact times will depend on the efficiency of the C compiler used as well as the computer).

up: table of contents previous section: 2. using cri-map with general pedigrees and disease loci

next section: 4. file structures