28 Oct 2010
SLINK: a general simulation program for linkage analysis
Based on an algorithm by Jurg Ott
Programming by Daniel E. Weeks with the help of Mark Lathrop
and Jurg Ott
Documentation by Daniel E. Weeks and Jurg Ott
**** WARNING ****
Familiarity with the LINKAGE programs is assumed. You will not
be able to correctly use these programs unless you are familiar
with the LINKAGE programs. This documentation does not attempt
to teach you how to use the LINKAGE programs. Furthermore, it is
assumed that you already have the LINKAGE package. If you do not
have it, it may be obtained from
ftp://linkage.rockefeller.edu/software/linkage/
However, we recommend using the faster version known as FASTLINK,
which is available from here:
http://www.ncbi.nlm.nih.gov/CBBresearch/Schaffer/fastlink.html
*****************************************************************
0. CURRENT VERSION . . . . . . . . . . . . . . . . . . . . . . 2
1. INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . 2
2. OVERVIEW OF SLINK . . . . . . . . . . . . . . . . . . . . . 3
3. INPUT FILES AND AVAILABILITY CODES. . . . . . . . . . . . . 4
4. HOW TO USE SLINK. . . . . . . . . . . . . . . . . . . . . . 6
5. TYPICAL APPLICATIONS. . . . . . . . . . . . . . . . . . . 11
5A. MSIM: Approximating the expected lod score. . . . . . . 11
5B. LSIM: Disease versus map of markers . . . . . . . . . . 14
5C. ISIM: Average maximum lod score and power . . . . . . . 17
5D. MSIM: Interpolating Zmax in each replicate. . . . . . . 19
5E. Finding the peak of the expected lod score curve. . . . 19
5F. MSIM and HOMOG and ELODHET: analysis under heterogeneity 19
6. TECHNICAL INFORMATION . . . . . . . . . . . . . . . . . . 20
6A. Program constants . . . . . . . . . . . . . . . . . 21
6B. References for SLINK. . . . . . . . . . . . . . . . 22
6C. Acknowledgments . . . . . . . . . . . . . . . . . . 22
7. LITERATURE. . . . . . . . . . . . . . . . . . . . . . . . 22
0. CURRENT VERSION
The current version of FastSLINK is 3.00 dated 28 Oct 2010.
FastSLINK is written in C and executes much more rapidly than the
original version of SLINK, which was written in Pascal. The
documentation below describes the original SLINK program, but is,
for the most part, applicable to the FastSLINK program. For
additional details regarding how FastSLINK differs from SLINK,
please see the README.txt file that is distributed with
FastSLINK.
Compared to previous versions, the following changes have been
implemented:
The default buffer size (128 bytes) for the input or output
files (whichever is usually larger) was increased to 4K bytes.
This resulted in the addition of at least one extra line of code
in nonstandard Pascal.
The critical levels for the maximum lod score (previously
fixed at 1, 2, and 3) are now user defined and must be furnished
to the analysis program in an input file called LIMIT.DAT (see
section 5). Similarly, input previously expected by SLINK on an
interactive basis must now be furnished in an input file called
SLINKIN.DAT (see section 3).
The quadratic interpolation routine, QUADMAX, in the MSIM
program has been rewritten, because it occasionally seemed to
give nonsensical results. It is now based on formulas (8.8)-
(8.14) in Ott (1991) and interpolates the maximum lod score given
three pairs of values (Z, theta). It does not carry out any extrapo-
lation beyond the smallest or largest theta value. If the lengths of
the two intervals differ by more than a factor of 4, no quadratic
interpolation is carried out, because the quadratic approximation
may not fit the lod score curve very well in those circumstances.
Also, no interpolation takes place when the first of the three
lod scores is less than -100 (at theta=0, presumably).
Known bugs: In the present version of SLINK, pedigree ID's
may be names or numbers. If they are numbers, these numbers must
not exceed the maximum number of pedigrees specified (it is
safest to number pedigrees consecutively starting with 1). This
is inconsistent with the LINKAGE programs, and the next program
version will correct this inconsistency.
1. INTRODUCTION
Simulation can provide approximate answers to questions that
are cumbersome or impossible to answer analytically. For exam-
ple, in a linkage study of a disease it is important to know if
the pedigrees you have collected (or plan to collect) will be
sufficient to detect linkage. The power to detect linkage
depends on a variety of factors, such as the structure of the
pedigree, the number of affected individuals, and the informa-
tiveness of the markers. The SLINK program described below
allows one to carry out such power calculations by simulating
genotypes at one locus given the phenotypes at another locus
linked with the first locus. If one is interested in simulating
under no linkage (e.g., to investigate exclusion possibilities or
p-values), the SLINK program may also be used but other programs
such as SIMULATE are in that case much more efficient than SLINK.
The SLINK package consists of a simulation program (SLINK)
and several analysis programs (MSIM, LSIM, ISIM, ELODHET) (this
allows analyzing the data under models different from those under
which the data were generated). As outlined below, the input
files to these programs conform to the rules required by the
LINKAGE programs.
SLINK implements a simulation algorithm developed by Jurg Ott
and described in:
1) Ott J (1989) Computer-simulation methods in human linkage
analysis. Proc Natl Acad Sci USA 86:4175-4178
The algorithm was implemented in the original SLINK computer
program package by Weeks, Ott, and Lathrop:
2) Weeks DE, Ott J, Lathrop GM (1990) SLINK: a general
simulation program for linkage analysis. Am J Hum Genet 47:A204
(abstr)
SLINK is based on the LINKAGE programs version 4.9 (Lathrop
et al., 1984) and accepts slightly modified LINKAGE data files as
explained in the file 'slink.txt'. The code has been updated to
be consistent with LINKAGE version 5.1.
The SLINK simulation program has been modified by Schaffer
and Weeks to use the algorithms developed by Cottingham et al:
3) Cottingham Jr RW, Idury RM, Schaffer AA (1993) Faster
sequential genetic linkage computations. Am J Hum Genet
53:252-263
Please cite references 1-3 if you use FastSLINK. Thank you.
Note that SLINK by itself is quite limited in terms of the
number of markers that it can handle. If you wish to simulate
large number of markers, then the SUP program will also be
necessary:
Lemire M. SUP: an extension to SLINK to allow a larger number
of marker loci to be simulated in pedigrees conditional on trait
values. BMC Genet. 2006 Jul 3;7:40. PubMed PMID: 16803631; PubMed
Central PMCID: PMC1524809.
SUP is available from the web site:
http://mlemire.freeshell.org/software.html
2. OVERVIEW OF SLINK
The documentation below describes the original SLINK program,
but is, for the most part, applicable to the FastSLINK program.
For additional details regarding how FastSLINK differs from
SLINK, please see the README.txt file that is distributed with
FastSLINK.
The program SLINK is a general computer simulation program
that employs a variation of the algorithm described by Ott
(1989). Suppose there are N people in a pedigree. Let x = (x1,
x2,...,xN) represent the vector of phenotypes of the N people in
the pedigree. Likewise, let g = (g1, g2,...,gN) represent the
vector of multi-locus genotypes including phase information. Then
the conditional probability distribution of the genotypes given
the phenotypes may be calculated by a series of successive risk
calculations:
P(g>=x) = P(g1>=x) P(g2>=g1,x) P(g3>=g1,g2,x)...
In the SLINK simulation algorithm, we calculate the conditional
probabilities (or risks) of all the possible multi-locus genotypes
with phase, P(g1>=x) and, based on these, we randomly assign one
of the genotypes to person 1. We then calculate P(g2>=g1,x),
taking into account the genotype g1 just generated, and randomly
assign a genotype to person 2, and so on. The process of succes-
sive risk calculations continues until all individuals in the
pedigree have been assigned a multi-locus genotype. This approach
is quite efficient since once a multi-locus genotype has been
assigned to one individual, it is known in all subsequent steps.
Note that this algorithm permits simulation conditional on any
combination of phenotypic data. For example, if the pedigree
were partially typed at a marker of interest, one could simulate
conditional on both the disease phenotypes and the marker data
currently available.
SLINK is based on the LINKAGE programs version 4.9 (Lathrop
et al., 1984) and accepts slightly modified LINKAGE data files as
explained below. The simulated pedigrees may be analyzed by
either the standard LINKAGE programs or by the special companion
programs (MSIM, ISIM, or LSIM), which have been modified to read
one replicate of the pedigrees at a time, rather than reading in
the entire pedigree file. In addition, the companion programs
provide some simple statistical summaries.
As described above, the data are simulated conditional on
the phenotypes given in the pedigree file, at the recombination
fraction(s) given in the data file. The loci are conceptually
divided into two categories: the 'trait' locus and the marker
loci; the trait locus need not be an affection-status locus type
but may be a codominant marker as well. There may be either one
or no trait locus, while there may be as many marker loci as you
desire. The trait locus is distinguished from the marker loci in
order to provide more flexibility in simulating the data, as
described below.
3. INPUT FILES AND AVAILABILITY CODES
SLINK requires three input files:
A) 'slinkin.dat' holding various parameter values, which
previously had to be furnished interactively. No text is permit-
ted between the values but text may follow the last input value
(see example file). The values are:
a) A seed for the random number generator. This
should be a different integer between 1 and 30,000 each
time. Larger numbers (> 25,000) are better.
b) The number of replicates desired.
c) The locus number identifying the trait locus.
This is the number of the trait locus in the 'simda-
ta.dat' file (see below). If you placed the trait
locus first in your 'simdata.dat' file, then the locus
number is 1. Input a zero if there is no trait locus,
i.e., if all the loci are markers.
d) The proportion of unlinked families (if you
want to allow for heterogeneity). Input 0 if you want
to simulate under the assumption of homogeneity (all
families are of the linked type, normal situation).
See section 5F for analyzing data generated under
heterogeneity.
B) 'simdata.dat', which is a standard LINKAGE data file in
MLINK-format.
C) 'simped.dat', which is a standard LINKAGE pedigree file
with an additional column inserted after the last phenotype
column. This additional column in the SLINK pedigree file
'simped.dat' contains the availability code, which controls what
type of phenotypes are written to the output file. These codes
are consistent with the codes used by the simulation program
SIMLINK (Boehnke, 1986; Ploughman and Boehnke, 1989).
Meaning of code for
Code Markers Trait
0 Unavailable Use orig. phenotypes as given in 'simped.dat'
1 Available Use simulated phenotypes
2 Available Use orig. phenotypes as given in 'simped.dat'
3 Unavailable Use simulated phenotypes
A person should be coded as "marker unavailable" if that
person will be assigned the phenotype "unknown" at each marker
locus in the output file 'pedfile.dat'. This is appropriate if
the person will not be typed for any markers because DNA samples
are not available from that person (i.e., they are dead or
uncooperative). Likewise, a person should be coded as "marker
available" if that person will be assigned simulated marker
phenotypes in the output file 'pedfile.dat'.
It is usually appropriate to "use original phenotypes" as
given in 'simped.dat' at the trait locus, because most simulation
studies of diseases are carried out conditional on the observed
phenotypes. If the "use original phenotypes" option is chosen,
then the trait phenotype remains as it was in the 'simped.dat'
input file. In rare cases, you may want to simulate trait
phenotypes by choosing the "use simulated phenotypes" option.
Keep in mind that these simulations are carried out condi-
tional on the phenotypes given in the 'simped.dat' input file.
Thus, if you want to indicate that someone is unavailable at the
trait locus, make their original phenotype unknown in 'sim-
ped.dat' and use one of the two "Use original phenotypes" codes.
If a simulation was being carried out using marker data
only, then only two availability codes are needed, as indicated
in the table below (codes 2 or 3 can still be used, if desired).
Every person who has an availability code of 0 will be given the
'unknown' phenotype at each marker locus. Every person who has
an availability code of 1 will be given a simulated phenotype at
each marker locus.
Code Meaning of code for Markers
0 Unavailable: Assign phenotype 'unknown' at each marker
1 Available: Assign simulated phenotypes at each marker
It is highly recommended that you read the example problems below
before attempting to use this simulation package.
EXAMPLE 1: Small 4-member family
Suppose we want to simulate the marker genotypes for two
affected children in a simple nuclear family where the two
parents must be carriers for a fully penetrant recessive disease,
and they are known to have four different marker alleles. We
first create a pedigree file called 'simpre.dat' which, after
processing it with MAKEPED (one of the programs of the LINKAGE
package), will become 'simped.dat' ('simpre.dat' is shown below;
the comments beginning with <= are not part of the file). We use
the availability code 2 for everyone, which means "Markers
available, Use original trait phenotypes". Thus, the 'pedfi-
le.dat' file (to be created by SLINK) will contain the original
trait phenotypes as given in 'simpre.dat' (and 'simped.dat'). At
the marker locus, person 1 will be 1/2 and person 2 will be 3/4,
since the simulations are carried out conditional on the pheno-
types given in the 'simped.dat' file. The two affected children
(persons 3 and 4) will be assigned simulated marker phenotypes,
since their original marker phenotypes in the 'simped.dat' file
are unknown (0 0).
File: simpre.dat
1 1 0 0 1 1 1 2 2 <= Father who is 1/2 at the marker
1 2 0 0 2 1 3 4 2 <= Mother who is 3/4 at the marker
1 3 1 2 1 2 0 0 2 <= First affected child
1 4 1 2 2 2 0 0 2 <= Second affected child
In the corresponding 'simdata.dat' file below, we give
penetrances corresponding to a fully penetrant recessive disease
at the trait locus and define a codominant four-allele system at
the marker locus.
File: simdata.dat
2 0 0 5 << NO. OF LOCI, RISK LOCUS, SEXLINKED (IF 1) PROGRAM
0 0.0 0.0 0 << MUT LOCUS, MUT MALE, MUT FEM, HAP FREQ (IF 1)
1 2
1 2 << AFFECTION, NO. OF ALLELES
0.9900 0.0100 << GENE FREQUENCIES Trait locus
1 << NO. OF LIABILITY CLASSES
0.0000 0.0000 1.0000 << PENETRANCES
3 4 << ALLELE NUMBERS, NO. OF ALLELES Marker locus
0.250000 0.250000 0.250000 0.250000 << GENE FREQUENCIES
0 0 << SEX DIFFERENCE, INTERFERENCE (IF 1 OR 2)
0.01000 << RECOMBINATION VALUES
1 0.01000 0.50000 << REC VARIED, INCREMENT, FINISHING VALUE
This example (files in subdirectory EXAMPLE1) is referred to
again in sections 5C and 5D.
4. HOW TO USE SLINK
First, compile the programs in the SLINK package to create
executables. Compilation instructions are in the documentation
file README.txt.
To carry out a simulation, proceed as follows (see also
figure 1, appended):
1) Create a 'simdata.dat' file defining the locus systems
and the true thetas.
2) Create a 'simped.dat' file defining the pedigree struc-
ture, phenotypes, and availability codes.
3) Create the 'slinkin.dat' file defining parameters for the
simulation (see section 3, above).
4) Run SLINK to create a 'pedfile.dat' with the simulated
data in it.
5) Create a 'datafile.dat' defining how the simulated data
should be analyzed.
6) Run UNKNOWN to create a 'speedfile.dat' and an 'ipedfi-
le.dat'.
7) Run the appropriate analysis program, such as MSIM or
LSIM.
These steps are described in detail below; the files of the
example data referred to below are in subdirectory EXAMPLE2.
Step 1) Create a 'simdata.dat' file defining the locus systems
and the true thetas.
The file 'simdata.dat' defines the locus systems and the
TRUE thetas (recombination fractions) under which the simulation
will be carried out. These true theta values are given in the
RECOMBINATION VALUES line; the subsequent line with increment
and finishing value is irrelevant for SLINK. The 'simdata.dat'
file must be in standard MLINK format and can be created with the
PREPLINK program of the LINKAGE package.
In this example, we are interested in simulating two linked
markers segregating down through two schizophrenia pedigrees,
conditional on the current schizophrenia diagnoses. This simula-
tion will be carried out under the assumptions that the locus
order is 599Ha---Schizophrenia---153Ra, with 8 cM between 599Ha
and the disease locus, and 5.7 cM between the disease locus and
153Ra.
File: 'simdata.dat'
3 0 0 5 << NO. OF LOCI, RISK LOCUS, SEXLINKED (IF 1) PROGRAM
0 0.0 0.0 0 << MUT LOCUS, MUT MALE, MUT FEM, HAP FREQ (IF 1)
1 2 3 << Order of loci
3 3 << ALLELE NUMBERS, NO. OF ALLELES Locus p105-599Ha/TaqI
0.3200 0.1600 0.5200 << GENE FREQUENCIES
1 2 << AFFECTION, NO. OF ALLELES Locus Schizophrenia
0.991500 0.008500 << GENE FREQUENCIES
3 << NO. OF LIABILITY CLASSES
1.0000 0.1400 0.0000 Normal
0.0000 0.8600 0.0000 Affected
1.0000 0.0000 0.0000 Married in Normals << PENETRANCES
3 2 << ALLELE NUMBERS, NO. OF ALLELES Locus p105-153Ra/XbaI
0.3300 0.6700 << GENE FREQUENCIES
0 0 << SEX DIFFERENCE, INTERFERENCE (IF 1 OR 2)
0.0739 0.0539 << RECOMB. VALUES Schiz. is 8 cM from p105-599Ha
2 0.01000 0.50000 << REC VARIED, INCREMENT, FINISHING VALUE
Step 2) Create a 'simped.dat' file defining the pedigree struc-
ture, phenotypes, and availability codes
First use a text editor to create a pedigree file 'simpre.dat'
analogous to the one shown above in example 1. Then use
MAKEPED to create the 'simped.dat' pedigree file.
The file 'simped.dat' defines the pedigrees to be simulated.
It is a LINKAGE pedigree file with an additional column contain-
ing the availability codes after the last phenotype field. All
simulations are carried out CONDITIONAL on the phenotypes given
in 'simped.dat'.
The example pedigree file contains two pedigrees from
Sherrington et al (1988), with marker locus 599Ha, the schizo-
phrenia locus, and marker locus 153Ra, in that order. Note that
all individuals are untyped at both markers. Also, most individ-
uals have an availability code of 2, indicating that all their
marker phenotypes will be simulated while their trait phenotypes
will be left as specified here. When a person is "marker avail-
able", marker phenotypes are simulated for ALL the marker loci;
it is not possible to make a person "marker available" for a
subset of the marker loci.
File: 'simped.dat'
1 1 0 0 3 0 0 1 1 0 0 2 3 0 0 2
1 2 0 0 3 0 0 2 0 0 0 2 2 0 0 0
1 3 1 2 0 4 4 2 0 0 0 2 2 0 0 2
1 4 1 2 0 5 5 2 0 0 0 2 1 0 0 2
1 5 1 2 0 6 6 1 0 0 0 2 2 0 0 2
1 6 1 2 0 7 7 2 0 0 0 2 1 0 0 2
1 7 1 2 0 8 8 1 0 0 0 2 1 0 0 2
1 8 1 2 0 9 9 1 0 0 0 2 2 0 0 2
1 9 1 2 0 10 10 1 0 0 0 2 2 0 0 2
1 10 1 2 14 12 12 2 0 0 0 2 2 0 0 2
1 11 0 0 14 0 0 1 0 0 0 2 3 0 0 2
1 12 1 2 0 13 13 1 0 0 0 2 2 0 0 2
1 13 1 2 0 0 0 2 0 0 0 2 1 0 0 2
1 14 11 10 0 15 15 2 0 0 0 2 1 0 0 2
1 15 11 10 0 16 16 1 0 0 0 2 2 0 0 2
1 16 11 10 0 17 17 1 0 0 0 2 2 0 0 2
1 17 11 10 0 0 0 2 0 0 0 2 1 0 0 2
2 1 0 0 3 0 0 1 1 0 0 2 3 0 0 0
2 2 0 0 3 0 0 2 0 0 0 2 2 0 0 2
2 3 1 2 0 4 4 2 0 0 0 2 2 0 0 2
2 4 1 2 0 5 5 1 0 0 0 2 2 0 0 2
2 5 1 2 0 6 6 1 0 0 0 2 2 0 0 2
2 6 1 2 10 8 8 1 0 0 0 2 2 0 0 2
2 7 0 0 10 0 0 2 0 0 0 2 3 0 0 2
2 8 1 2 13 0 0 2 0 0 0 2 2 0 0 2
2 9 0 0 13 0 0 1 0 0 0 2 3 0 0 2
2 10 6 7 0 11 11 1 0 0 0 2 1 0 0 2
2 11 6 7 0 12 12 1 0 0 0 2 1 0 0 2
2 12 6 7 0 0 0 1 0 0 0 2 2 0 0 2
2 13 9 8 0 14 14 1 0 0 0 2 1 0 0 2
2 14 9 8 0 0 0 1 0 0 0 2 2 0 0 2
Step 3) Create the 'slinkin.dat' input file for SLINK holding the
parameters defined in section 3, above. Please note that the
random number seed should be changed for every new simulation.
Step 4) Run SLINK to create a 'pedfile.dat' containing the
simulated data. As outlined in section 3, SLINK requires several
parameters in an input file, 'slinkin.dat'. The input files for
SLINK are 'simdata.dat' and 'simped.dat' (see the flow chart in
figure 1).
SLINK outputs the simulated data in the file 'pedfile.dat',
while the parameters (as described above) and the thetas are
written to 'simout.dat'. If one of the companion analysis
programs is run, it will put the information from the most recent
'simout.dat' into the output file it creates (figure 1). Note
that the 'pedfile.dat' is a bona-fide LINKAGE pedigree file and
can be analyzed by any of the regular LINKAGE programs. However,
it may contain many replicates of the pedigrees (one set after
another), so that in most cases it is more practical to use the
modified versions of the LINKAGE programs (MSIM, ISIM, LSIM) that
are designed to analyze the data a replicate at a time.
In our example, we asked for simulated marker data while
maintaining the original trait phenotypes (Availability code 2)
for all but two people. Note that the two people with availabil-
ity code 0 have unknown marker genotypes.
File: 'pedfile.dat' (first replicate of both families)
1 1 0 0 3 0 0 1 1 1 3 2 3 1 2 2 Linked Mk Avail; Tr orig
1 2 0 0 3 0 0 2 0 0 0 2 2 0 0 0 Linked Mk Unkno; Tr orig
1 3 1 2 0 4 4 2 0 1 3 2 2 1 2 2 Linked Mk Avail; Tr orig
1 4 1 2 0 5 5 2 0 1 3 2 1 1 1 2 Linked Mk Avail; Tr orig
1 5 1 2 0 6 6 1 0 1 3 2 2 1 2 2 Linked Mk Avail; Tr orig
1 6 1 2 0 7 7 2 0 3 3 2 1 2 2 2 Linked Mk Avail; Tr orig
1 7 1 2 0 8 8 1 0 3 3 2 1 2 2 2 Linked Mk Avail; Tr orig
1 8 1 2 0 9 9 1 0 1 1 2 2 1 1 2 Linked Mk Avail; Tr orig
1 9 1 2 0 10 10 1 0 1 3 2 2 1 2 2 Linked Mk Avail; Tr orig
1 10 1 2 14 12 12 2 0 1 1 2 2 1 1 2 Linked Mk Avail; Tr orig
1 11 0 0 14 0 0 1 0 1 3 2 3 2 2 2 Linked Mk Avail; Tr orig
1 12 1 2 0 13 13 1 0 1 1 2 2 2 1 2 Linked Mk Avail; Tr orig
1 13 1 2 0 0 0 2 0 3 3 2 1 1 2 2 Linked Mk Avail; Tr orig
1 14 11 10 0 15 15 2 0 1 3 2 1 1 2 2 Linked Mk Avail; Tr orig
1 15 11 10 0 16 16 1 0 1 1 2 2 2 1 2 Linked Mk Avail; Tr orig
1 16 11 10 0 17 17 1 0 1 3 2 2 1 2 2 Linked Mk Avail; Tr orig
1 17 11 10 0 0 0 2 0 1 3 2 1 1 2 2 Linked Mk Avail; Tr orig
2 1 0 0 3 0 0 1 1 0 0 2 3 0 0 0 Linked Mk Unkno; Tr orig
2 2 0 0 3 0 0 2 0 1 2 2 2 2 2 2 Linked Mk Avail; Tr orig
2 3 1 2 0 4 4 2 0 1 1 2 2 2 2 2 Linked Mk Avail; Tr orig
2 4 1 2 0 5 5 1 0 1 1 2 2 2 2 2 Linked Mk Avail; Tr orig
2 5 1 2 0 6 6 1 0 1 1 2 2 2 2 2 Linked Mk Avail; Tr orig
2 6 1 2 10 8 8 1 0 1 2 2 2 2 2 2 Linked Mk Avail; Tr orig
2 7 0 0 10 0 0 2 0 3 3 2 3 1 2 2 Linked Mk Avail; Tr orig
2 8 1 2 13 0 0 2 0 1 2 2 2 2 2 2 Linked Mk Avail; Tr orig
2 9 0 0 13 0 0 1 0 1 3 2 3 2 1 2 Linked Mk Avail; Tr orig
2 10 6 7 0 11 11 1 0 2 3 2 1 2 1 2 Linked Mk Avail; Tr orig
2 11 6 7 0 12 12 1 0 2 3 2 1 2 2 2 Linked Mk Avail; Tr orig
2 12 6 7 0 0 0 1 0 1 3 2 2 2 1 2 Linked Mk Avail; Tr orig
2 13 9 8 0 14 14 1 0 1 2 2 1 2 2 2 Linked Mk Avail; Tr orig
2 14 9 8 0 0 0 1 0 1 1 2 2 2 2 2 Linked Mk Avail; Tr orig
Step 5) Create a 'datafile.dat' defining how the simulated data
should be analyzed.
The analysis programs require five input files (figure 1).
Two are made by UNKNOWN ('ipedfile.dat' and 'speedfile.dat', see
step 6). One is made by SLINK ('simout.dat', see step 4). The
fourth, 'datafile.dat', must be made using PREPLINK prior to
running the desired analysis program. The fifth input file,
'limit.dat', contains three threshold values (e.g., 1 2 3) used
to determine the proportion of replicates exceeding a given lod
score limit.
The 'datafile.dat' is a standard LINKAGE data file which
determines how the analyses of the simulated data will be carried
out. It must be in MLINK format for MSIM; ILINK format for ISIM;
and LINKMAP format for LSIM.
The example 'datafile.dat' below is in MLINK format and is thus
appropriate for input into MSIM (the modified version of MLINK).
File: 'datafile.dat'
3 0 0 5 << NO. OF LOCI, RISK LOCUS, SEXLINKED (IF 1) PROGRAM
0 0.0 0.0 0 << MUT LOCUS, MUT MALE, MUT FEM, HAP FREQ (IF 1)
1 3 2
3 3 << ALLELE NUMBERS, NO. OF ALLELES Locus p105-599Ha/TaqI
0.320000 0.160000 0.520000 << GENE FREQUENCIES
1 2 << AFFECTION, NO. OF ALLELES Locus Schizophrenia
0.991500 0.008500 << GENE FREQUENCIES
3 << NO. OF LIABILITY CLASSES
1.0000 0.1400 0.0000 Normal
0.0000 0.8600 0.0000 Affected
1.0000 0.0000 0.0000 Married in Normals << PENETRANCES
3 2 << ALLELE NUMBERS, NO. OF ALLELES Locus p105-153Ra/XbaI
0.330000 0.670000 << GENE FREQUENCIES
0 0 << SEX DIFFERENCE, INTERFERENCE (IF 1 OR 2)
0.11970 0.12500 << RECOMBINATION VALUES
2 0.12500 0.38000 << REC VARIED, INCREMENT, FINISHING VALUE
NOTE: If you choose the increment size too small, you will get
an error message that maxpnt is too small. Maxpnt is the number
of theta's (or map positions in LSIM) at which the lod score is
evaluated. You may fix this problem by either choosing a larger
increment or by increasing maxpnt and recompiling.
Step 6) Run UNKNOWN to create a 'speedfile.dat' and an 'ipedfi-
le.dat'
Once the 'pedfile.dat' has been created by SLINK, then it is
necessary to process this file with the program UNKNOWN (of the
LINKAGE package) before running any of the analysis programs.
UNKNOWN creates the pedigree file 'ipedfile.dat' and the 'speed-
file.dat' ('speedfil.dat' on DOS machines). These two files are
needed for input into MSIM, ISIM, or LSIM.
Step 7) Run the appropriate analysis program, such as MSIM or
LSIM.
5. TYPICAL APPLICATIONS
The analysis programs, MSIM, LSIM, and ISIM, each require
three input files:
1) A parameter file called 'limit.dat'. This file holds
three thresholds/limits for the maximum lod score; the programs
will approximate the probability with which the maximum lod score
exceeds each of the three thresholds. Typical threshold values
are 1, 2, and 3. Note that numbers must have at least one digit
to the left of any decimal point, for example, 0.5; a number
given as .5 would lead to an error.
2) A locus file, 'datafile.dat', which is analogous to the
one used for the LINKAGE programs. You may simply copy 'simda-
ta.dat' to 'datafile.dat' and modify the 'datafile.dat' file to
correspond (i) to the analysis program to be used and (ii) to
reflect the theta values at which analysis is to be carried out.
Note that 'datafile.dat' is also used as an input file to the UN-
KNOWN program.
3) A pedigree file, 'ipedfile.dat', created by the UNKNOWN
program.
Details of programs usage are given below.
5A. MSIM: Approximating the expected lod score
First we show how to use MSIM, which summarizes its results
in the file 'msim.dat'. The first part of 'msim.dat' contains
information defining the simulation, such as the random number
seed, the number of replicates, the requested proportion of
unlinked families, and the trait locus number. This information
is taken from the most recent 'simout.dat'. Note that the thetas
and the locus order presented in the 'simout.dat' section pertain
to the model under which the simulation was carried out (as
previously specified in the 'simdata.dat' file). These may
differ from the thetas and locus order used in the analysis of
the simulated data. The rest of 'msim.dat' provides statistical
information.
If two loci are used, then the results are reported on the
traditional lod score scale. When three or more loci are used,
multi-point lod scores are computed as the log likelihood at the
current thetas minus the log likelihood with the theta involving
the trait locus set to 0.50. For this reason, when MLINK or MSIM
are used, the trait locus must be either the leftmost or right-
most locus for these statistics to make sense. Normally, MSIM,
like MLINK, will be used for analyses involving only two loci.
However, in this example, we use three loci, but place the trait
locus (number 2) on the right by specifying the locus order to be
'1 3 2' in the 'datafile.dat' file (see above). Lod scores for
the position of a disease versus a map of markers are usually
computed with the LINKMAP (LSIM, see next section); MSIM is used
here for demonstration purposes.
File: 'msim.dat'
********* Data from most recent SIMOUT.DAT *********
The random number seed is: 25432
The number of replications is: 20
The requested proportion of unlinked families is: 0.000
The trait locus is locus number: 2
Summary Statistics about simped.dat
Number of pedigrees 2
Number of people 31
Number of females 12
Number of males 19
There were 2 in category: Marker Unknown; Trait original
There were 0 in category: Marker Available; Trait simulated
There were 29 in category: Marker Available; Trait original
There were 0 in category: Marker Unknown; Trait simulated
LINKAGE (V4.91) WITH 3-POINT AUTOSOMAL DATA
-----------------------------------
LINKED ORDER OF LOCI: 1 2 3
-----------------------------------
-----------------------------------
TRUE THETAS FOR LINKED ORDER 0.073900 0.053900
-----------------------------------
-----------------------------------
UNLINKED ORDER OF LOCI: 2 1 3
-----------------------------------
-----------------------------------
TRUE THETAS FOR UNLINKED ORDER 0.500000 0.119834
-----------------------------------
Elapsed Time for one replicate = 8 seconds
Elapsed Time = 162 seconds or 2.70 minutes.
-----------------------------------
Actual proportion of unlinked families: 0.000
********* End of most recent SIMOUT.DAT *********
ORDER OF LOCI: 1 3 2
Average Multipoint Lod Scores at Given Thetas
Number of replicates = 20
----------------------------------------------------------------
THETAS 0.120 0.125
----------------------------------------------------------------
Pedigree Average StdDev Min Max
1 0.928233 0.823127 -0.499552 2.372393
2 0.581052 0.458913 -0.066598 1.564117
Study 1.509285 0.725744 0.163738 3.215271
----------------------------------------------------------------
THETAS 0.120 0.250
----------------------------------------------------------------
Pedigree Average StdDev Min Max
1 0.673438 0.524425 -0.170536 1.607466
2 0.355322 0.371757 -0.404606 0.957419
Study 1.028760 0.507228 -0.259767 2.132345
----------------------------------------------------------------
THETAS 0.120 0.375
----------------------------------------------------------------
Pedigree Average StdDev Min Max
1 0.313556 0.231926 -0.036782 0.742686
2 0.140380 0.209244 -0.369045 0.398341
Study 0.453935 0.265833 -0.275562 0.937106
----------------------------------------------------------------
Brief explanation of the output:
The file 'msim.dat' contains information defining the
simulation, followed by some tables providing statistical infor-
mation about the distribution of the simulated lod scores. The
'Average' column provides the expected (or mean) lod score by
pedigree and by study (i.e., set of families). The 'StdDev'
column provides the standard deviation of the lod score. The
'Min' column lists the minimum lod score encountered in all the
replicates, and the 'Max' column lists the maximum lod score
encountered in all the replicates (and when the pedigrees are
used together in one study). Note that the only column in which
the Study value will equal the sum of the Pedigree values is the
'Average' column. If one is interested in approaching the
absolute smallest (or absolute largest) lod score in the whole
study, one should add up the min. (or max.) values (if all have
the same sign) over pedigrees rather than looking at the Study
min. (or max.) value.
5B. LSIM: Disease versus map of markers
If we want to run LSIM (the modified version of LINKMAP), we
need a different 'datafile.dat' than the one used above for MSIM
because the data file for LSIM must be in standard LINKMAP
format. Like LINKMAP, LSIM is most appropriate for calculating
lod scores of the trait locus at various positions across a fixed
map of marker loci. As mentioned above, a multi-point lod score
requires calculation of the likelihood of the data with the trait
"off" the map first. Thus, there are two requirements for
getting accurate results out of LSIM: 1) the trait locus must
start out as the leftmost locus; and 2) the trait locus must be
placed "off" the map on the left at a recombination fraction of
0.50. This allows calculation of multi-point lod scores as the
trait locus is moved across the whole map. In this example, the
trait locus is number 2 and so the locus order must be specified
as '2 1 3' in order to have the trait locus be off the map on the
left. Since the trait locus must be placed off the map, the
first recombination fraction is 0.50.
File: 'datafile.dat'
3 0 0 4 << NO. OF LOCI, RISK LOCUS, SEXLINKED (IF 1) PROGRAM
0 0.0 0.0 0 << MUT LOCUS, MUT RATE, HAPLOTYPE FREQUENCIES (IF 1)
2 1 3
3 3 << ALLELE NUMBERS, NO. OF ALLELES Locus p105-599Ha/TaqI
0.320000 0.160000 0.520000 << GENE FREQUENCIES
1 2 << AFFECTION, NO. OF ALLELES Locus Schizophrenia
0.991500 0.008500 << GENE FREQUENCIES
3 << NO. OF LIABILITY CLASSES
1.0000 0.1400 0.0000 Normal
0.0000 0.8600 0.0000 Affected
1.0000 0.0000 0.0000 Married in Normals << PENETRANCES
3 2 << ALLELE NUMBERS, NO. OF ALLELES Locus p105-153Ra/XbaI
0.330000 0.670000 << GENE FREQUENCIES
0 0 << SEX DIFFERENCE, INTERFERENCE (IF 1 OR 2)
0.50000 0.11980 << RECOMBINATION VALUES
2 0.11970 4 << LOCUS VARIED, FINISHING VALUE, NU OF EVALUATIONS
LSIM (the modified version of LINKMAP) creates the following
output file, which has been edited to conserve space. Note how
LSIM, unlike LINKMAP, automatically moves the trait locus across
each interval in the fixed map of marker loci, as indicated by
the different locus orders in the tables below (the middle number
on the last line of the 'datafile.dat' is irrelevant for LSIM).
File: 'lsim.dat' (edited)
********* Data from most recent SIMOUT.DAT *********
-----------------------------------
LINKED ORDER OF LOCI: 1 2 3
-----------------------------------
-----------------------------------
TRUE THETAS FOR LINKED ORDER 0.073900 0.053900
-----------------------------------
-----------------------------------
UNLINKED ORDER OF LOCI: 2 1 3
-----------------------------------
-----------------------------------
TRUE THETAS FOR UNLINKED ORDER 0.500000 0.119834
-----------------------------------
Actual proportion of unlinked families: 0.000
********* End of most recent SIMOUT.DAT *********
Average Multipoint Lod Scores at Given Thetas
Number of replicates = 20
----------------------------------------------------------------
Locus Order: 2 1 3 THETAS 0.500 0.120
Number of replicates with a maximum at this location 0
----------------------------------------------------------------
Pedigree Average StdDev Min Max
1 0.000000 0.000000 0.000000 0.000000
2 0.000000 0.000000 0.000000 0.000000
Study 0.000000 0.000000 0.000000 0.000000
----------------------------------------------------------------
Locus Order: 2 1 3 THETAS 0.125 0.120
Number of replicates with a maximum at this location 1
----------------------------------------------------------------
Pedigree Average StdDev Min Max
1 0.986943 0.957538 -0.563674 2.811669
2 0.553123 0.587166 -0.495850 1.583681
Study 1.540066 0.943955 -1.059524 3.268075
----------------------------------------------------------------
Locus Order: 1 2 3 THETAS 0.060 0.068
Number of replicates with a maximum at this location 2
----------------------------------------------------------------
Pedigree Average StdDev Min Max
1 1.147581 1.148037 -0.900799 3.205506
2 0.751585 0.636813 -0.343115 2.097035
Study 1.899166 1.040807 -0.212925 4.120526
----------------------------------------------------------------
Locus Order: 1 2 3 THETAS 0.090 0.037
Number of replicates with a maximum at this location 0
----------------------------------------------------------------
Pedigree Average StdDev Min Max
1 1.097943 1.163443 -1.264877 3.126892
2 0.733534 0.603973 -0.154687 2.095361
Study 1.831477 1.055195 -0.426049 4.131400
----------------------------------------------------------------
Locus Order: 1 3 2 THETAS 0.120 0.125
Number of replicates with a maximum at this location 3
----------------------------------------------------------------
Pedigree Average StdDev Min Max
1 0.928120 0.823000 -0.499557 2.372007
2 0.580943 0.458917 -0.066556 1.564077
Study 1.509064 0.725652 0.163398 3.214886
----------------------------------------------------------------
Locus Order: 1 3 2 THETAS 0.120 0.250
Number of replicates with a maximum at this location 0
----------------------------------------------------------------
Pedigree Average StdDev Min Max
1 0.673346 0.524329 -0.170544 1.607176
2 0.355246 0.371761 -0.404694 0.957392
Study 1.028592 0.507163 -0.259854 2.132040
----------------------------------------------------------------
Locus Order: 1 3 2 THETAS 0.120 0.375
Number of replicates with a maximum at this location 0
----------------------------------------------------------------
Pedigree Average StdDev Min Max
1 0.313506 0.231874 -0.036785 0.742532
2 0.140356 0.209227 -0.369050 0.398359
Study 0.453862 0.265785 -0.275485 0.936943
----------------------------------------------------------------
The output in LSIM.DAT is very similar to that found in
MSIM.DAT (Average, Standard Deviation, Minimum and Maximum).
However, there is an additional line that records the number of
replicates with a maximum at the location. This number gives a
measure of how often the trait locus would be mapped to that
particular location, providing a sense of the best supported
location for the trait locus. In our example, we simulated with
the trait locus between the two marker loci, and we find that 3
out of 20 times, the trait would have mapped into the interval at
a theta of .120 from the first marker locus.
NOTE: The program parameter maxpnt, which defines the
maximum number of points at which the lod score can be evaluated,
may cause problems when the increment of theta is too small.
Maxpnt appears in MSIM and LSIM, but not in ISIM. If one sets
the increment of theta too small, MSIM or LSIM may terminate with
the following error message: ERROR: maxpnt is too small. The
solution is to 1) evaluate lod score at less points or 2) set
maxpnt higher and recompile the programs. In LSIM, this may mean
choosing points evaluated in each interval rather than over the
whole map.
5C. ISIM: Average maximum lod score and power
With ISIM (as with ILINK), the recombination fraction
between two loci can be iterated and, therefore, the lod score
obtained for a given replicate is maximized over the theta
values. Thus, ISIM can approximate the average (expectation) of
the maximum lod score.
In our example application of ISIM, we use the files from
Example 1, where we simulated the marker genotypes for two
affected children in a simple nuclear family where the two
parents are known to have four different marker alleles. The
'simdata.dat' and 'simped.dat' have been explained above. The
'datafile.dat' must be in ILINK format. If the recombination
fraction between the two loci is iterated, then the average of
the maximum lod score will be calculated. If the recombination
fraction is not iterated, then the average of the lod score
(unmaximized) will be calculated at the requested thetas. In the
file below, we request that the recombination fraction be iterat-
ed. Note: ISIM may not maximize the log likelihood over theta
correctly if the disease gene frequencies are extreme (even when
they are fixed). You may also want to designate locus 1 or locus
2 as the locus with iterated parameters.
File: 'datafile.dat'
2 0 0 3 << NO. OF LOCI, RISK LOCUS, SEXLINKED (IF 1) PROGRAM
0 0.0 0.0 0 << MUT LOCUS, MUT MALE, MUT FEM, HAP FREQ (IF 1)
1 2
1 2 << AFFECTION, NO. OF ALLELES
0.9900 0.0100 << GENE FREQUENCIES
1 << NO. OF LIABILITY CLASSES
0.0000 0.0000 1.0000 << PENETRANCES
3 4 << ALLELE NUMBERS, NO. OF ALLELES
0.250000 0.250000 0.250000 0.250000 << GENE FREQUENCIES
0 0 << SEX DIFFERENCE, INTERFERENCE (IF 1 OR 2)
0.01000 << RECOMBINATION VALUES
1 << THIS LOCUS MAY HAVE ITERATED PARS
1
When we run ISIM, we obtain the following output, which
reports the average of the maximum lod score:
File 'isim.dat' (edited):
********* Data from most recent SIMOUT.DAT *********
The random number seed is: 29876
The number of replications is: 200
The requested proportion of unlinked families is: 0.000
The trait locus is locus number: 1
Summary Statistics about simped.dat
Number of pedigrees 1
Number of people 4
Number of females 2
Number of males 2
There were 0 in category: Marker Unknown; Trait original
There were 0 in category: Marker Available; Trait simulated
There were 4 in category: Marker Available; Trait original
There were 0 in category: Marker Unknown; Trait simulated
LINKAGE (V4.91) WITH 2-POINT AUTOSOMAL DATA
-----------------------------------
LINKED ORDER OF LOCI: 1 2
-----------------------------------
-----------------------------------
TRUE THETAS FOR LINKED ORDER 0.010000
-----------------------------------
Actual proportion of unlinked families: 0.000
********* End of most recent SIMOUT.DAT *********
Average Maximum Lod Score
The number of replicates is 200
--------------------------------------------------------
Average Maximum | StdDev | Min | Max
0.576825 0.117744 0.000000 0.600859
--------------------------------------------------------
--------------------------------------------------------
Number of maximum lod scores greater than
a given constant
--------------------------------------------------------
Constant Number Percent
1 161 80.500 (assumed values)
2 0 0.000
3 0 0.000
--------------------------------------------------------
Note that there is no theta value reported with the Average
Maximum Lod Score. This is because each maximum lod score (in a
replicate) may correspond to a different estimate of theta.
Also, note that the average maximum lod score is not additive
over families or studies. The average lod score (ELOD), calcu-
lated previously, does possess this desirable property of additi-
vity (see Ott, 1991).
The last table appearing in 'isim.dat' gives the number and
percentage of replicates in which the maximized lod score (Zmax)
was found greater than the given constants. The proportion of
replicates with Zmax > 3 is an approximation to the power, i.e., to
finding a significant linkage.
5D. MSIM: Interpolating Zmax in each replicate
The average of the maximum lod score (calculated by ISIM)
may be much more time consuming to compute, because maximization
of the lod score requires many evaluations of the lod score. To
partially avoid this problem, MSIM can estimate maximum lod
scores by quadratic interpolation. The quadratic interpolation
option works automatically, but only in the special case where
there are two loci, autosomal inheritance, and the 'datafile.dat'
for MSIM requests evaluation of the lod score at three or more
distinct points. To get the best results from the quadratic
interpolation, you should evaluate the lod score at more than
three points. Also, it is better not to choose theta = 0.0 as
one of the points, because if there is an obligate recombination,
then the lod score is minus infinity (use theta = 0.00001 or
0.001 instead). For example, if we run the problem above,
starting theta at 0.01 and increasing it in increments of 0.10 up
to 0.41, for a total of 5 points, we get the results seen below,
which are very similar to those obtained above by ISIM:
From the file: 'msim.dat'
----------------------------------------------------------------
Average Maximum Lod Scores based on quadratic interpolation
----------------------------------------------------------------
Pedigree Average StdDev Min Max
1 0.579848 0.118361 0.000000 0.604008
----------------------------------------------------------------
5E. Finding the peak of the expected lod score curve
Theoretically, the expected lod score should have its
maximum (the ELOD) at the true recombination fraction. In a
simulation, one may verify this by computing the average lod
score at several assumed theta values in the vicinity of the true
recombination fraction under which the data are generated. The
peak should occur at or near the true recombination fraction.
MSIM or LSIM may be used for these calculations.
5F. MSIM and HOMOG and ELODHET: analysis under heterogeneity
As outlined above, SLINK can generate data under heterogene-
ity (alpha <1) but one should not analyze data so generated assuming
homogeneity, for example, using output from the MSIM program. If
families generated with alpha <1 are analyzed with MSIM, that is,
assuming alpha=1, the resulting lod scores are too low and do not
correspond to the ELOD under heterogeneity; also, the estimates
of theta are biased upwards.
When generating data under heterogeneity, the easiest
solution for analysis is to focus on the expected lod score
(ELOD); working with the probability that the maximum lod score
exceeds a given threshold requires more complicated manipulations
than described here. The replicates generated by SLINK are
analyzed with MSIM but the results produced by MSIM (MSIM.DAT)
are disregarded. Instead, one takes the LODFILE.DAT output
generated by MSIM and analyzes it using either the HOMOG program
(version 3.30 or higher) or a utility program called ELODHET.
The latter program uses LODFILE.DAT and MSIM.DAT as input files
and computes for each family its expected lod score under hetero-
geneity. Presently, its use is restricted to two loci (a trait
and a marker locus). For analysis by the HOMOG program, the data
in LODFILE.DAT will have to be completed to form a proper input
file for HOMOG.
When data are generated under heterogeneity, it is usually
most appropriate to also analyze them under heterogeneity. The
ELODHET program accomplishes this task. It carries out two types
of analyses:
1) For the given values of a (proportion of families with
linkage) and q (recombination fraction in families with linkage)
under which the family data were generated, ELODHET calculates
expected lod scores Z for each family and, for all families
jointly, the probability (power) that the lod score at a and q
exceeds given thresholds (as furnished in the 'limit.dat' file).
The expected lod score under heterogeneity is defined as Z(a, q)
= log10[L(a, q)] - log10[L(1, 0.5)], where L is the usual
likelihood under heterogeneity, that is, for the i-th family,
Li(a, q) = a Li(q) + 1 - a, and Li(q) is the antilog of the i-th
family's lod score.
2) Expected maximum lod scores are calculated by evaluating
in each replicate, for all families jointly, the maximum
likelihood over a grid of values of a and q (irrespective of the
a and q values used to generate the data). That is, each
simulated replicate furnishes Zmax(a, q) = log10[Lmax(a, q)] -
log10[L(0, 0.5)], where a and q are estimates obtained in a given
replicate. For the Zmax values, their average is computed and the
proportion of replicates in which Zmax exceeds given thresholds
(power). Note that Zmax involves two degrees of freedom, and the
associated lod score threshold is thus somewhat less conservative
than that for the lod score under homogeneity. In this sense, the
Zmax under heterogeneity is comparable to obtaining separate
estimates for male and female recombination fractions.
To use the ELODHET program, one must first run SLINK and
MSIM. For the latter run, lod scores should be evaluated at
suitable intervals, for example, starting at q = 0.00001 and
increasing in steps of 0.05 with a limiting value of 0.47. In its
present implementation, the step size for a in the analysis is
taken to be 0.025. As in the HOMOG programs, no interpolation is
carried out to approximate maximum lod scores. The ELODHET
program reads the LODFILE.DAT file created by MSIM and furnishes
its results in an output file called ELODHET.DAT.
6. TECHNICAL INFORMATION
As indicated below, the programs in this package are modi-
fied versions of LINKAGE programs. They were modified by Daniel
E. Weeks with help from Mark Lathrop. Please address any ques-
tions or problems to:
Daniel E. Weeks
University of Pittsburgh
Department of Human Genetics
130 DeSoto Street
A300 Crabtree Hall
Pittsburgh, PA 15261
(412) 624-5388 or (412) 624-3018
FAX: (412) 624-3020
Email: weeks@pitt.edu
The programs are distributed by J. Ott in versions for Vax
(VMS) computers ("Disk" 21) and microcomputers running under DOS
or OS/2 ("Disks" 20a through d). The DOS and OS/2 versions have
been adapted to Prospero Pascal. For details on compiling and
linking, see the PROSPERO.TXT file. Disk 20b contains a running
version of the UNKNOWN program. You may also use the UNKNOWN
program furnished with the LINKAGE programs.
Differences between Prospero and Vax Pascal (double preci-
sion real variables):
Prospero Vax
------------- --------------
Top line none [G_FLOATING]
Type real= LONGREAL DOUBLE
Closing files CLOSE(ff,true) CLOSE(ff)
Clock function explicit (present) built-in
File buffer FBUFFER(ff,4096) not used
Assign statements ASSIGN not used
In addition, in procedure inib, replace ASSIGN by OPEN for Vax
Pascal.
6A. Program constants
Program Purpose/Description
---------------------------------------------------
Simulation SLINK General simulation program
Analysis MSIM Modified version of MLINK
ISIM Modified version of ILINK
LSIM Modified version of LINKMAP
The analysis programs, MSIM, ISIM, and LSIM, read the number
of pedigrees per replicate (i.e., study) from the summary file
'simout.dat' produced by SLINK.
There are several constants that may be changed to modify
the behavior of the programs. This requires editing the source
code followed by recompilation. Compilation instructions are in
the documentation file README.txt.
A standard version of one of the LINKAGE programs, such as
MLINK, produces a lot of output. For the purposes of conserving
disk space and speeding up the SLINK programs, most of the
'standard' output has been turned off by a logical flag. This
constant is called 'prn' or 'print' and is identified in the
source code of each program by a comment. Set the constant equal
to TRUE if you want more output.
For an affection status locus, the unaffected individuals
are indicated by the integer constant assigned to 'unaff'.
In the LINKAGE programs, the phenotype at a quantitative
locus may consist of several measurements (or traits). However,
SLINK supports only ONE measurement (or trait) for a quantitative
locus. Unaffected males at a sex-linked locus are indicated by
the integer constant assigned to 'unafqu'.
MSIM will write a list of lod scores by pedigree to the file
'lodfile.dat' if the constant 'lodprint' is set to TRUE. Howev-
er, it only writes out the male thetas. This may be useful if
you wish to produce a histogram of lod scores.
6B. References for SLINK
Please use the following two references when reporting
results based on SLINK:
Ott J (1989) Computer-simulation methods in human linkage
analysis. Proc Natl Acad Sci USA 86:4175-4178
Weeks DE, Ott J, Lathrop GM (1990) SLINK: a general simula-
tion program for linkage analysis. Am J Hum Genet 47:A204
(abstr)
6C. Acknowledgments
We would like to thank Joseph Terwilliger and Suzanne Leal
for their helpful comments. Support by the W.M. Keck Foundation
and grants MH44292 from NIMH and HG00008 from the National Center
for Human Genome Research are gratefully acknowledged.
7. LITERATURE
Boehnke M (1986) Estimating the power of a proposed linkage
study: a practical computer simulation approach. Am J Hum Genet
39:513-527
Lange K, Matthysse S (1989) Simulation of pedigree genotypes
by random walks. Am J Hum Genet 45:959-970
Lathrop GM, Lalouel JM, Julier C, Ott J (1984) Strategies
for multilocus linkage analysis in humans. Proc Natl Acad Sci
USA 81:3443-3446
Ott J (1985) Analysis of human genetic linkage, first
edition. Johns Hopkins University Press, Baltimore
Ott J (1989) Computer-simulation methods in human linkage
analysis. Proc Natl Acad Sci USA 86:4175-4178
Ott J (1991) Analysis of Human Genetic Linkage, second
edition. Johns Hopkins University Press, Baltimore and London
Ploughman LM, Boehnke M (1989) Estimating the power of a
proposed linkage study for a complex genetic trait. Am J Hum
Genet 44:543-551
Sandkuyl L, Ott J (1989) Determining informativity of marker
typing for genetic counseling in a pedigree. Hum Genet 82:159-
162
Sherrington R, Brynjolfsson J, Petursson H, Potter M,
Dudleston K, Barraclough B, Wasmuth J, Dobbs M, Gurling H (1988)
Localization of a susceptibility locus for schizophrenia on
chromosome 5. Nature 336:164-167
Weeks DE, Ott J, Lathrop GM (1990) SLINK: a general simula-
tion program for linkage analysis. Am J Hum Genet 47:A204 (abstr)
Figure 1:
simdata.dat simped.dat
| |
| |
| |
| |
\/ \/
==============================
slinkin.dat->= SLINK = ----> simout.dat
==============================
| |
| |
| |
| |
\/ |
|
datafile.dat pedfile.dat |
| | |
| | |
\/ \/ |
========================== |
= UNKNOWN = |
========================== |
| | |
| | |
| | |
\/ \/ |
datafile.dat ipedfile.dat speedfile.dat |
| | | |
| | | |
| | | |
\/ \/ \/ |
============================== |
limit.dat->= MSIM, ISIM, LSIM = <-------------|
| ============================== |
| | | |
| | | |
| | | |
| \/ \/ |
| |
| outfile.dat msim.dat |
| lodfile.dat isim.dat <---------------|
| | lsim.dat
| | |
\/ \/ \/
==============================
= ELODHET =
==============================
|
|
\/
elodhet.dat