TABLE.DOC

                           Documentation for TABLE.EXE

        This is a very simple program which takes the output from MLINK 
        or LINKMAP and tabulates it in a more easily digestible form. It 
        can also produce files suitable for graphing lod scores and for 
        input to HOMOG, BTEST and FASTMAP. 

        The input file is the OUTPUT.DAT file which the LINKAGE programs 
        produce. When using LINKSYS this file is renamed to be        
        FILENAME.RES. With LCP or the LINKSYS program MCP it is possible 
        to run a whole load of analyses all at once and output them into 
        one results file. If you do this TABLE will not work correctly - 
        it expects to find the output from only one analysis by either 
        LINKMAP or MLINK. If you have set up a batch file to run several 
        analyses at once, you must edit the output so that the result of 
        each analysis is in a different .RES file. TABLE has actually only 
        been tested with the results files produced by LINKSYS v. 4.11 and 
        LINKAGE v. 4.7 and 5. With LINKMAP it assumes that the test locus 
        is named first and that the whole map has been produced. With 
        MLINK it assumes that the "lod score table" output option has been 
        selected. TABLE may work in other circumstances - I just haven't 
        tried it. 

        If you are working with more than one pedigree at a time then 
        make sure that the "byfamily" switch in the LINKAGE programs is 
        set to true so that likelihood values are output for each family 
        - it will not increase the time that it takes the program to 
        run. Otherwise only the total lod scores across pedigrees will 
        be available. 

        If your results have ended up in FILENAME.RES then to use TABLE 
        just enter:

        TABLE FILENAME

        A new file will be created on your disk called FILENAME.TAB. 
        This contains a table of lod scores at different recombination 
        frequencies, or with LINKMAP at different distances taking 
        distance 0 to be at 0% recombination with the leftmost fixed 
        marker. The lod scores are calculated by subtracting the log10 
        likelihood at theta=50% from the log10 likelihood at other 
        recombination fractions. Reombination fractions are converted to 
        distances in centimorgans using the Kosambi mapping function.

        There are six optional switches which can be used with TABLE 
        to produce other files: 
        
        TABLE FILENAME /G /H /A[val] /B /I /X     (MSDOS)

        table filename -g -h -a[val] -b -i -x     (Unix)

        If /G is selected a file called FILENAME.GRP will be created. 
        This provides the input for a Shareware graphing program 
        available from me called EASIGRAF, which forms part of the 
        EASISTAT package. To see the graph of lod scores one simply 
        enters: 
        
        EASIGRAF FILENAME.GRP
        
        If /X is selected a file called FILENAME.XGR will be created. 
        This is a graph file for the ACE/gr graphing program, which 
        runs on workstations and terminals using X. Run the 
        program (called xvgr or xmgr) and select the "Read sets" command 
        from the "File" menu. Read in the file (FILENAME.XGR). Click on 
        the autoscale button (AS) and the graph of lod scores should be 
        displayed. ACE/gr was written by Paul Turner and is available in 
        source form from ftp.ccalmr.ogi.edu in CCALMR/pub/acegr.
        
        If /H is selected a file called FILENAME.HOM is produced which 
        can be read in by Jurg Ott's HOMOG program to perform the A-test 
        of homogeneity of linkage. (See Ott, 1985 Analysis of Human 
        Genetic Linkage, John Hopkins University Press and Ott, Ann Hum 
        Genet 47:311-320, 1983). A version of the program is available 
        from Jurg Ott but unfortunately this no longer uses the same 
        format as the listing in the book so some editing would be 
        necessary. The source code for a version transcribed into C 
        which uses the original format is included together with the 
        executable. 

        If /A is selected then the TABLE program itself performs the A-
        test calculations. It outputs the adjusted lod score (log 10 of 
        the likelihood ratio) obtained under the assumption of 
        heterogeneity, called A-lod, together with alpha, the fraction of 
        linked families which produce this maximum lod. It does not 
        explicitly perform the test for statistical significance which the 
        original HOMOG program does. To do this by hand, subtract the 
        maximum conventional (total) lod from the maximum value of A-lod 
        and multiply by 2ln(10) (twice the natural log of 10, about 
        4.605). According to Ott's original account, this may be taken as 
        a chi-squared statistic with one degree of freedom providing 
        evidence to support the hypothesis of heterogeneity if linkage is 
        assumed. It is also possible to supply a fixed value for alpha, in 
        which case the adjusted lod will be calculated assuming that the 
        given fraction of families is linked (otherwise the TABLE program 
        searches for the value of alpha which yields the maximum lod 
        score). To supply a fixed value for alpha, enter it directly after 
        the /A switch, e.g. /A0.5, /A0.8. Any number of fixed alpha 
        values can be specified, as well as or instead of the /A switch on 
        its own. The adjusted lod scores are appended to others in the 
        table and graph files. However for reasons of clarity the graph 
        file written for EASIGRAF will not be set up to display these 
        adjusted lod scores by default. In order to have them displayed 
        then either the graph files should be edited (to insert the 
        relevant DATATYPE:XY lines) or else the Data menu can be used 
        from within EASIGRAF.
        
        If /I is selected a file called FILENAME.INP is created which 
        contains recombination fraction and lod score pairs as needed 
        for input into the FASTMAP program. This produces estimates of 
        multipoint lod scores from two-point scores, and has been 
        submitted for publication to Human Heredity. It is available 
        from me or (preferably) from the gene-server ftp site or one of 
        its mirrors (anonymous ftp from ftp.bchs.uh.edu and change 
        directory to /pub/gene-server/dos, or by email from gene-
        server@bchs.uh.edu or gene-server%bchs.uh.edu@CUNYVM, send mail 
        with subject: SEND DOS HELP). Normally, only recombination 
        fractions between 0.01 and 0.4 are output. It probably isn't 
        advisable to change this, but if you do want to try it you can 
        change the values for MIN_RF_TO_FIT and MAX_RF_TO_FIT in TABLE.C 
        and recompile it. 

        If /B is selected a file called FILENAME.BTE is produced which 
        can be input into Neil Risch's BTEST program for homogeneity of 
        linkage. (See Risch, Am J Hum Genet, 42:353-364, 1988). Copies 
        of this program are available from Neil Risch at Yale University 
        School of Medicine, Department of Epidemiology and Public 
        Health, 60 College Street, PO Box 3333, New Haven, CT 06510; 
        (203) 785-2838; BITNET: RISCH@YALEMED. 

        Note that to produce input for HOMOG and BTEST the .RES file 
        must (obviously) contain output from more than one pedigree and 
        the likelihoods by family must be available. For BTEST the 
        output must be from MLINK, and not LINKMAP. 

        I hope people find this useful.


        Dave Curtis          September 1993
        

        Academic Department of Psychiatry
        St Mary's Hospital Medical School
        Praed Street
        London W2 1NY, England                   Phone:  071 725 1993

        Janet:       dcurtis@UK.AC.CRC
        Elsewhere:   dcurtis@CRC.AC.UK
        EARN/Bitnet: dcurtis%CRC@UKACRL
        Usenet: ...!mcsun!ukc!mrccrc!D.Curtis