Unphased
From Vd
Contents |
UNPHASED
General description
UNPHASED is a suite of programs for association analysis of multilocus haplotypes from unphased genotype data. Unphased v.3 can be run using the executable unphased in /site/bin.
Note: Previous version of unphased, were split into several executables. We also have these executables available. These are:
- tdtphase: TDT and HHRR analysis for nuclear families
- cocaphase: case/control data
- qtphase: quantitative trait in unrelateds
- pdtphase: pedigree disequilibrium test
- qpdtphase: quantitative trait PDT
Application home page
http://www.mrc-bsu.cam.ac.uk/personal/frank/software/unphased/
Proper citation
Dudbridge F (2003) Pedigree disequilibrium tests for multilocus haplotypes. Genet Epidemiol 25:115-21
Quick start
You only need one input file (ped-file), optionally with a linkage datafile (all programs) and a haplotype frequ ency file (PDTPHASE and QPDTPHASE) - see below.
Input file format
All programs take a linkage pre-makeped pedigree file as input. This is a plain text file with lines of the foll owing whitespace-delimited format:
PedID SubID PID MID Sex Status Allele1a Allele1b Allele2a Allele2b ....
Although the UNPHASED programs will run regardless of Mendelian errors or missing parental genotypes, it is stro ngly recommend that you run your file through PEDCHECK (http://watson.hgen.pitt.edu/register/, needs registratio n before download) before analysis. A linkage datafile can be used to define the marker names (following "#" as in Genehunter) and to allow differen t pedigree file formats to the above. PDTPHASE and QPDTPHASE can optionally use an input file containing haplotype frequencies. Each line of the file specifies one haplotype. The alleles are contained in the first n columns (n being the length of the haplotype), followed by the frequency. For example,
1 1 0.25 1 2 0.25 2 1 0.2 2 2 0.3
If a haplotype is omitted from the frequency file, it is assumed to have zero frequency. Useful commands
The basic command line is:
program <inputfile> [options]
The name of the inputfile is added by the system. Here are all the available options: they are not available in all programs, if a program does not implement an option, it will print a list of the options it does implement.
- EM: EM estimation of uncertain haplotype frequencies; if omitted, only phase-certain haplotypes are used
- LD: displays D', r^2 and Cramer's V when two-locus haplotypes are selected (ie. "-window 2")
- bygenotype: conditions on genotypes rather than on haplotypes at the conditioning loci
- cellcounts: determines rare haplotypes on the basis of expected counts rather than frequencies
- chrX: identifies data from chromosome X; male alleles in the first column for each marker
- compare: specify first haplotype in a pairwise comparison; wild cards may be specified by "0"
- condition: select the conditioning markers
- conditiontype: specify a conditioning haplotype; the null is that all haplotypes containing this sub-haplotype have same risk
- datafile: specify a linkage datafile in Genehunter format
- droprare: rare haplotypes are grouped together, but under the null are not grouped with other haplotypes, so th e global test has one fewer degree of freedom. The test has greater power when the effect is in common haplotype s.
- epsilon: specify a convergence threshold for the likelihood in the EM algorithm; default is 1e-8 which is quite strict
- freqfile: specify a file containing fixed haplotype frequencies
- hpt: homozygous parent test; uses only the parents which are homozygous in the conditioning loci. Optionally s pecify which haplotype to be homozygous for, using the -conditiontype option
- individual: perform tests of individual haplotypes by grouping all others together
- maineffects: test main effects only, in conditional testing
- marker: select markers to analyse from the pedigree file
- missing: estimate missing genotypes by looping over allele values which are consistent with the data
- nointeraction: no interaction terms at all; test and conditioning haplotypes are both modelled with main effect s only
- observed: analyse only the haplotypes which are observed with certainty in at least one subject
- onesib: use only the first informative sibling from each family
- output: write most of the output to a file instead of the screen
- paraff: select parental affection status; 1=unaffected, 2=affected
- parsex: select parental sex; 1=male, 2=female
- pedfile: specify the pedigree file (not required as an option)
- permutation: specify number of replicates for permutation test. Estimates the significance of the best result correcting for multiple testing of haplotypes, loci and phenot ypes.
- phenotype: select the phenotypes for analysis, requires a datafile. If omitted, the first phenotype in the pedigree file is used. Multiple phenotypes can be listed.
- rare: specify frequency below which haplotypes are grouped as rare. Should be between 0 and 1, but if greater than 1, interpreted as a percentage unless -cellcounts is spec ified. Maximum-likelihood frequencies are re-computed after identification of rare haplotypes.
- rarein: followed by "both" (default), "either", "case" or "control", selects which frequency is used to determine whether a haplotype is rare
- reference: specify reference haplotype which has odds ratio 1
- restarts: restart the EM algorithm from multiple random starting points
- robustperm: use the same randomised transmission status for all sibs, therefore robust to prior linkage
- sibaff: select sibling affection status
- sibsex: select sibling sex
- sum: compute the "PDT-sum" statistic (see Martin et al. AJHG 68:1065). By default, "PDT-avg" is computed
- window: specify the length of haplotypes; the selected markers will be analysed in groups of this size - see examples below
- with: specify second haplotype in a pairwise comparison; no wildcards allowed
- zero: specify frequency below which haplotype frequencies are trimmed to zero. Such haplotypes are assumed to be absent in the population when calculating degrees of freedo m. The default is 1e-8.