RAREFY COMMAND

From Genome Analysis Wiki
Jump to: navigation, search

Useful Wiki Pages

RAREFY COMMANDS

Options:
  Input Files : --ped [], --dat []
      Methods : --MCMC, --traitIncreasing, --traitDecreasing
        Trait : --inverseNormal, --useCovariates, --traitName []
   Parameters : --maf [1.0e-03], --effect [1.00]
         MCMC : --seed, --chains [3], --iterations [50000000]
        Other : --famList [], --famID [], --cpus [5], --prefix []

Input Files

--ped [string]

followed by the file name of your MERLIN format PED file.

--dat [string]

followed by the file name of your MERLIN format DAT file.

Methods

--MCMC [on|off]

default is off, but will be automatically turned on for families larger than 25 individuals.

--traitIncreasing [on|off]

default is off; can be used together with --traitDecreasing options or used alone.

--traitDecreasing [on|off]

default is off; can be used together with --traitIncreasing options or used alone.

Trait

--inverseNormal [on|off]

default is off; on will inverse normalize (quantile normalize) traits before fitting variance component model or adjusting covariates (if requested by --useCovariates option).

--useCovariates [on|off]

default is off; on will adjust any covariates saved in PED/DAT file (covariates are labeled with "C" in DAT file).

--traitName [string]

default is to analyze all traits saved in PED/DAT file (traits are labeled with "T" in DAT file).

Parameters

--maf [double]

default is 0.001; this option let you change the value of the fixed prior frequency of a trait-associated variant of interest.

--effect [double]

default is 1 SD; this option let you change the value of the fixed prior effect size of a trait-associated variant of interest.

MCMC

--seed [int]

--chains [int]

The default is 3. This options allows users to specify number of MCMC chains to run with different starting point. In general, less than 10 chains is recommended. The result is averaged over all chains.

--iterations [int]

The default is 50 million iterations. This option allows users to specify number of total iterations run in MCMC. Each chain will have the same amount of iterations divided from the total.

Others

--famList [string]

This options allows analysis of only a subset of families listed in a file. However, all families included in PED/DAT file will be used to fit variance component model.

--famID [string]

This option allows analysis of only a single family. However, all families included in PED/DAT file will be used to fit variance component model.

--cpus [int]

This option allows parallel computing using multiple CPUs. However, the maximum number of CPUs used is capped by number of chains specified in --chains option, if MCMC method is used. NOTE: a version of RAREFY allowing parallel computing must be compiled to use this option. Please refer to COMPILE for more information.

--prefix [string]

This option allows users to customize the output prefix generated by RAREFY.