Single Nucleotide Polymorphism Spectral Decomposition Lite (SNPSpDlite)
SNPSpDlite is not designed to analyse >2000 SNPs genotyped in 1000's of individuals.
Firstly, unless you are analysing less common variants (e.g., MAF below 0.05) then 100 individuals will be sufficient to estimate LD, so you should only upload a subset of your data.
Secondly, the SNPSpD estimates were published in 2004 and 2005, and there has since been some major advances, both in accuracy and efficiency. Therefore, if you wish to estimate the effective number of independent markers on larger datasets then I recommend
i) my matSpDlite approach (which only requires users to upload a correlation matrix),
ii) download the matSpDlite.R R script to perform matSpDlite analysis on your local machine, or
iii) use the Genetic Type I error calculator (GEC).
SNPSpDlite takes the same input at SNPSpD, but ONLY outputs the LD correlation matrix, Meff and MeffLi estimates and GOLD input (i.e., does not perform time-consuming varimax/promax rotations) thus allowing users to obtain Meff/MeffLi values for large numbers (1000s) of SNPs.
Users may also be interested in the related site (SNPSpDsuperlite) which is an even leaner version of SNPSpD.
Users may also be interested in the related site (matSpD) which only requires users to upload a correlation matrix.
Please use the following reference when
reporting results based on SNPSpD:
Nyholt DR (2004) A simple correction for multiple testing for SNPs in linkage disequilibrium with each other. Am J Hum Genet 74(4):765-769.
To run SNPSpD using all fully genotyped
To run SNPSpD using all family members:
|Tel: +61-7-3362 0258||Find Us|
|Email: daleN@qimr.edu.au||Contact Us|