SKAT test

SNP-set (Sequence) Kernel Association Test Method

1. Details

1.1 Command interface

vtools show test SKAT

Name:          SKAT
Description:   SKAT (Wu et al 2011) wrapper of its original R implementation
usage: vtools associate --method SKAT [-h] [--name NAME]
                                      [-k {linear,linear.weighted,IBS,IBS.weighted,quadratic,2wayIX}]
                                      [--beta_param BETA_PARAM BETA_PARAM]
                                      [-m {davies,liu,liu.mod,optimal}]
                                      [-i {fixed,random}]
                                      [--logistic_weights PARAM PARAM]
                                      [-r [CORR [CORR ...]]]
                                      [--missing_cutoff MISSING_CUTOFF]
                                      [--resampling N] [--small_sample]
                                      [--resampling_kurtosis N]
                                      {quantitative,disease}

SNP-set (Sequence) Kernel Association Test (Wu et al 2011). This is a wrapper
for the R package "SKAT" implemented & maintained by Dr. Seunggeun Lee, with a
similar interface and minimal descriptions based on the SKAT package
documentation (May 11, 2012). Please refer to
http://http://cran.r-project.org/web/packages/SKAT/ for details of usage. To
use this test you should have R installed with SKAT v0.75 or higher. The SKAT
commands applied to the data will be recorded and saved in the project log
file.

positional arguments:
  {quantitative,disease}
                        Phenotype is quantitative trait or disease trait (0/1
                        coding). Default set to "quantitative"

optional arguments:
  -h, --help            show this help message and exit
  --name NAME           Name of the test that will be appended to names of
                        output fields, usually used to differentiate output of
                        different tests, or the same test with different
                        parameters.
  -k {linear,linear.weighted,IBS,IBS.weighted,quadratic,2wayIX}, --kernel {linear,linear.weighted,IBS,IBS.weighted,quadratic,2wayIX}
                        A type of kernel. Default set to "linear.weighted".
                        Please refer to SKAT documentation for details.
  --beta_param BETA_PARAM BETA_PARAM
                        Parameters for beta weights. It is only used with
                        weighted kernels. Default set to (1,25). Please refer
                        to SKAT documentation for details.
  -m {davies,liu,liu.mod,optimal}, --method {davies,liu,liu.mod,optimal}
                        A method to compute the p-value. Default set to
                        "davies". Please refer to SKAT documentation for
                        details.
  -i {fixed,random}, --impute {fixed,random}
                        A method to impute missing genotypes. Default set to
                        "fixed". Please refer to SKAT documentation for
                        details.
  --logistic_weights PARAM PARAM
                        This option, if specified, will get the logistic
                        weights from genotype matrix Z and apply this weight
                        to SKAT. It requires two input parameters par1 and
                        par2. To use the SKAT default setting, type
                        `--logistic_weights 0.07 150'. Please refer to SKAT
                        documentation for details.
  -r [CORR [CORR ...]], --corr [CORR [CORR ...]]
                        The pho parameter of SKAT test. Default is 0. Please
                        refer to SKAT documentation for details.
  --missing_cutoff MISSING_CUTOFF
                        a cutoff of the missing rates of SNPs. Any SNPs with
                        missing rates higher than cutoff will be excluded from
                        the analysis. Default set to 0.15
  --resampling N        Number of resampling using bootstrap method. Set it to
                        '0' if you do not want to apply resampling.
  --small_sample        This option, if evoked, will apply small sample
                        adjustment "SKAT_Null_Model_MomentAdjust" for small
                        sample size and binary trait. Please refer to SKAT
                        documentation for details.
  --resampling_kurtosis N
                        Number of resampling to estimate kurtosis, for small
                        sample size adjustment. Set it to '0' if you do not
                        wnat to apply the adjustment. The SKAT default setting
                        is 10000. Please refer to SKAT documentation for
                        details.

1.2 Application

vtools associate rare status -m "SKAT --name skat quantitative" --group_by refGene.name2 --\
to_db skat -j8 > skat.txt

INFO: 3180 samples are found
INFO: 2632 groups are found
INFO: Starting 8 processes to load genotypes
Loading genotypes: 100% [=========================================================================================================================================] 3,180 32.8/s in 00:01:36
Testing for association: 100% [================================================================================================================================] 2,632/147 8.9/s in 00:04:56
INFO: Association tests on 2632 groups have completed. 147 failed.
INFO: Using annotation DB skat in project test.
INFO: Annotation database used to record results of association tests. Created on Wed, 30 Jan 2013 21:34:23


vtools show fields | grep skat

skat.refGene_name2           refGene_name2
skat.sample_size_skat        Sample size
skat.Q_stats_skat            Test statistic for SKAT, "Q"
skat.pvalue_skat             p-value


head skat.txt

refGene_name2	sample_size_skat	Q_stats_skat	pvalue_skat
AADACL4	3180	33707.7	0.379148
ABCD3	3180	1178.25	0.961708
AAMP	3180	5905.71	0.612598
ABCB10	3180	55121.9	0.109206
ABCB6	3180	16500.2	0.812062
ABCG5	3180	9829.17	0.76832
ABI2	3180	42491.9	0.0100467
ABHD1	3180	1315.49	0.880286
ABL2	3180	794.385	0.963097