------------------------- SETTING UP THE PROGRAM ------------------------- This file is 0README.TXT - from /pub/academic/biology/dna-mutations/hyperg The files in this subdirectory hold the program for analysis of mutational spectra data that is described in Carcinogenesis, 15, 2281-2285, 1994. . The files in this directory MUST BE transferred using FTP as BINARY files. The files are compressed to speed up transfer time via FTP. The file ending in "EXE" is a self-extracting archives. This means when you execute them (by typing their names) some files will "pop out", and you will have some files present after uncompression. There are two files in /pub/academic/biology/dna-mutations/hyperg: 0README.TXT - this file. HG_UNPAK.EXE - compressed, self-extracting archive. If you got these files via the INTERNET using FTP and they don't work properly (typically the HG_UNPAK.EXE will not uncompress), you probably did not transfer the files correctly. These files are BINARY and must be transferred as such. Please see your system manager for details about binary FTP, I cannot help you at my end with your file transfers. When you download the program from a mainframe to your PC, remember to SET FILE TYPE TO BINARY if you are using KERMIT. Whatever you are using to get the file from the mainframe to the PC, it must be done as a binary file transfer. May I suggest: 1) Make a directory on the hard disk called HYPERG (use the DOS MD command). 2) Change the current directory to C:\HYPERG by entering: CD C:\HYPERG 3) Copy all files to the C:\HYPERG subdirectory (use the DOS COPY command). 4) Unpack the compressed files by entering: HG_UNPAK 5) The following files should be present after uncompressing: 00READ.ME - this file HG-PUBL.C - C source code IDM2.C - C source code HG-PUBL.EXE - executable program SAMPLE.DAT - a sample data file 6) After decompressing the file HG_UNPAK.EXE, it should be deleted from the hard disk. Please keep a copy of the compressed files on floppy disks in case something bad happens. ----------------------- RUNNING THE PROGRAM ----------------------- The program yields a p value, a p value of <= 0.05 indicates that the spectra are most likely not drawn from the same population. This program makes use of random numbers, therefore, you will probably NOT get exactly the same answer if you run the program twice with the same input spectra. The answers should be close, but they may not be identical. It would be bad idea to run this program hundreds of times trying to get the p value to fall below 0.05 just once, and then to report the 0.05 p value. Remember those lectures about integrity in science? The algorithm implemented here is described in WT Adams and TR Skopek, J Mol Biol, 194, 1987. This algorithm is currently viewed as one of the best for mutational spectra calculations; for a detailed statistical analysis, see WW Piegorsch and AJ Bailer, Genetics, 136, 403-416, 1994. You should set up a table with the data from spectra 1 in one column and the data from spectra 2 in the other column. The columns contain the number of mutations of a particular NATURE for each base pair. For example, you may have the following two spectra: bp wt specific spectra spectra position base mutation 1 2 ------- ---- -------- -------- -------- 44 A A->C 0 1 A->G 0 1 A->T 0 0 45 C C->A 0 0 C->G 0 0 C->T 0 0 46 G G->A 3 1 G->C 0 1 G->T 1 1 Rows with no mutants (0 0) are discarded and the final two-column table is set up, like this. 0 1 0 1 3 1 0 1 1 1 This table now contains the two input spectra. The file containing these two columns of numbers is then submitted to the program. The numbers must be separated by one or more spaces, or by one or more tabs. Non-numeric data (this is, alphabetic characters) is NOT ALLOWED. Rows with no information (that is, "0 0" are likewise NOT ALLOWED). This program does alot of error checking, so most files with incorrect data will be caught before the program runs. Please note that the input file MUST BE A NORMAL ASCII FILE. If you create the input file with a modern word processor, make sure you save it as TEXT ONLY. Do not save the file in the native word processing format. Do not save the file as ".DOC" if you are using Word, or ".WP" if you are using WordPerfect. The program then generates a number of random tables that contains the same number of mutants in each row as the input spectra. The number of iterations that you request is really the number of random tables to generate. 1700 random tables gives fairly good precision. The input table is compared to each random table. If the input spectra is "more unusual" than the random table, this is scored as a hit. The p value is "hits / (total number of random tables)". The maximum number of samples (that is, the total of column 1 and column 2) that are allowed is 1000. However, since the source code is provided, a researcher can increase this limit and recompile the program, if desired. At any rate, first you create the input data file. Then you start the program by typing: HG-PUBL The program will prompt you for the file name with your data and will then run the program. To run the sample spectra provided with the program, just enter the file name, which is: SAMPLE.DAT If you have any questions, please EMAIL me. Digitally, Neal Cariello cariello@sunsite.unc.edu