On this page... (hide)

  1. 1. Introduction
  2. 2. Details
    1. 2.1 Command interface
    2. 2.2 Example

1.  Introduction

vtools_report meta_analysis implements meta-analysis methods detailed in Willer et al 20101. Two statistics are available: the sample size based statistic and inverse variance based statistic. Input of this command are multiple text or compressed text files of association results delimited by tabs.

2.  Details

2.1  Command interface

% vtools_report meta_analysis -h
usage: vtools_report meta_analysis [-h] [--beta col] [--pval col] [--se col]
                                   [-n col] [--link col [col ...]] [-m method]
                                   [--to_db database] [-v {0,1,2}]
                                   file [file ...]

positional arguments:
  file                  Input text files in the format of $vtools associate
                        output (supports plain text, gz or bz2 compressed text
                        files)

optional arguments:
  -h, --help            show this help message and exit
  --beta col            column number of beta
  --pval col            column number of p-value
  --se col              column number of standard error
  -n col, --size col    column number of sample size
  --link col [col ...]  columns that links entries of two datasets
  -m method, --method method
                        Method (choose from "ssb" for sample based method and
                        "ivb" for inverse variance based method), default set
                        to "ssb"
  --to_db database      will write the results also to a sqlite3 database
                        compatible with vtools associate result format
  -v {0,1,2}, --verbosity {0,1,2}
                        Output error and warning (0), info (1) and debug (2)
                        information of vtools and vtools_report. Debug
                        information are always recorded in project and
                        vtools_report log files.

2.2  Example

Input files are

% zless study1.gz
refGene_name2   sample_size_VT  num_variants_VT total_mac_VT    beta_x_VT       pvalue_VT       std_error_VT    num_permutations_VT     MAF_threshold_VT
A1BG    159     2       24      0.0191151       0.891109        0.213265        1000    0.00628931
A1BG-AS1        309     2       43      -0.409899       0.563437        0.27416 1000    0.00161812
A1CF    327     6       10      0.274825        0.275724        0.18807 1000    0.00152905
A2M     330     6       21      -0.0123306      0.985015        0.153102        1000    0.00151515
A2ML1   331     8       19      -0.280548       0.25974 0.159701        1000    0.00151057
...

study2.gz is in the same format. The effect size is the 5th column, p-value the 6th column and sample size 1st column. The two data-sets are matched by gene name, i.e., the first column. The meta analysis command is:

% vtools_report meta_analysis study1.gz study2.gz --beta 5 --pval 6 --size 2 --link 1 \
% --to_db study1and2 > study1and2.txt

Result are outputted to both text file and database.

% less study1and2.txt
refGene_name2	p_meta	sample_size_meta	beta_x_VT_1	pvalue_VT_1	sample_size_VT_1	beta_x_VT_2	pvalue_VT_2	sample_size_VT_2
RESP18	6.380E-01	238	0.570054	0.1998	119	-0.228113	0.537463	119
STX6	8.583E-01	633	0.010475	0.875125	302	-0.434853	0.691309	331
ARFGEF2	7.984E-01	645	0.161718	0.417582	314	-0.169275	0.251748	331
TSEN34	3.459E-01	104	0.604153	0.323676	52	0.461471	0.729271	52
...

To load the meta analysis result to the project

% vtools use study1and2.DB --linked_by name2
INFO: Using annotation DB study1and2 in project SSc.
INFO: Combined association tests result database. Created on Thu, 14 Mar 2013 17:35:44
INFO: 7459 out of 23242 refGene.name2 are annotated through annotation database study1and2
 

1 C. J. Willer, Y. Li and G. R. Abecasis (2010). METAL: fast and efficient meta-analysis of genomewide association scans. Bioinformatics