MetaSKAT

Association Analysis with the Meta SKAT R Program

The R extension for RTest is available HERE. For details on the format of this script please refer to the RTest method documentation.

1. Example

We analyze association with a binary trait named X6 conditioning on 3 covariates X8, X9, X10. Group information is provided by race. For each testing group, data is first cleaned by removing samples missing greater than 50% calls, then by removing variants missing 50% calls.

1.1 Basic command

The following command uses the default parameter settings in the R script:

vtools associate variant X6 --covariates X8 X9 X10 race \
       -m "RTest ~/MetaSKAT.VAT.R --name MetaSKAT --out_type 'D' --group_colname 'race' " \
       --discard_samples '%(NA)>0.5' --discard_variants '%(NA)>0.5' \
       --group_by 'refGene.name2' -j8 --to_db MetaSKAT > MetaSKAT.txt

INFO: Loading R script '/home/gw/MetaSKAT.VAT.R'
INFO: 252 samples are found
INFO: 131 groups are found
INFO: Starting 8 processes to load genotypes
Loading genotypes: 100% [===================================] 252 72.5/s in 00:00:03
Testing for association: 100% [==============================] 131/122 10.4/s in 00:00:12
INFO: Association tests on 131 groups have completed. 122 failed.
INFO: Using annotation DB MetaSKAT in project mskat.
INFO: Annotation database used to record results of association tests. Created on Tue, 16 Apr 2013 01:40:03
INFO: 131 out of 23242 refgene.name2 are annotated through annotation database MetaSKAT

Note that with such stringent cleaning criteria only 9 out of 131 groups are analyzed. In the R program we throw an error on having no samples, only one variant, or only one study group. Please take a look at the *.log file for details of the errors.

refGene_name2	pvalue_MetaSKAT	n_pop1_MetaSKAT	n_pop2_MetaSKAT	n_pop3_MetaSKAT
gene1	0.686969	79	84	NAN
gene2	0.0697923	84	68	NAN
gene3	0.0684316	81	83	NAN
gene4	0.426611	63	84	NAN
gene5	0.591128	70	77	NAN
gene6	0.00586012	84	81	NAN
gene7	0.306343	77	84	NAN
gene8	0.0951658	81	84	NAN
gene9	0.0988938	84	84	84

1.2 Use other parameters

From the way MetaSKAT.VAT.R is written we known out_type and group_colname are required. They have to be explicitly specified at all times. Other parameters have default values, which can be altered by passing them like, for example:

... -m "RTest ~/MetaSKAT.VAT.R --name MetaSKAT \
   --out_type 'D' --group_colname 'race' --pval.method 'davies' " ...