fastMETA — Usage
Quick start
-
Prepare a tab-separated file with at least:
variable, BETA1, SE1, BETA2, SE2. You may add more BETA/SE pairs (e.g.BETA3, SE3). - Go to the Tool tab and upload your file.
- Choose a Method and (if using method 2/3) a Heterogeneity estimator .
- Click Run fastMETA
- When finished, inspect the Results table and click Download results to save a copy.
Input file structure
Tab-separated with at least these columns (you may include more than two BETA/SE pairs):
| variable | BETA1 | BETA2 | SE1 | SE2 |
|---|---|---|---|---|
| var1 | 0.0715430377 | 0.0218272071 | 0.0220505147 | 0.0220699109 |
| var1 | 0.0966807914 | 0.0854040595 | 0.0604912844 | 0.0607718594 |
| var1 | 0.1383389831 | 0.0450079846 | 0.0998447463 | 0.1000947719 |
| var2 | 0.1298049745 | 0.1219084187 | 0.0261241204 | 0.0261034413 |
| var2 | 0.0737540951 | 0.0285241921 | 0.0219609195 | 0.0219791685 |
| var3 | 0.0912965958 | 0.0744658800 | 0.0605097372 | 0.0607870714 |
| var3 | 0.1513790485 | 0.0512020952 | 0.0978128888 | 0.0980551991 |
Methods overview
- method 1 — xmeta Marginal Method of moments (MMoM) het_est ignored . Suitable for quick scans or homogeneous settings.
-
method 2 — per-variable correlation coefficient
: estimates correlation separately for each variable. Requires selecting
het_est. -
method 3 — shared correlation coefficient
: assumes a common correlation across variables. Requires selecting
het_est.
Output columns (typical)
-
variable: identifier as provided in the input. -
BETA_META: pooled effect estimate. -
SE_META: standard error of the pooled effect. -
Wald: Wald statistic ≈ BETA_META/SE_META . -
P: two-sided p-value.
About fastMETA
fastMETA is developed for efficient multivariate meta-analysis of GWAS.
- Correspondence: Pantelis G. Bagos (pbagos@compgen.org)
- Main developer: Georgios A. Manios (gmanios@uth.gr)
- GitHub: https://github.com/pbagos/fastMETA
- Computational Genetics Group: https://sites.google.com/compgen.org/lab/home?authuser=0
