I need to get information regarding the best "free software" that can be used for meta-analysis of association studies dealing with single nucleotide polymorhisms and disease susceptibility?
RevMan from Cocrane Collaboration is free and easier to use than R. R requires a lot of training while RevMan is quite simple and straight-forward and there is on-line training possibilities.
As Fares says, Excel is also an option, but it requires that you do a lot of manual work for each analysis and this gets complicated when you have more complicated statistical assumptions.
A little late, perhaps, but I feel tempted to advertise my own software: http://www.ncbi.nlm.nih.gov/pubmed/22971100
You can download a linux-binary, here: http://yamas.meb.uni-bonn.de/
In case you encounter problems to set the it up, you can, of-course, contact me. It offers two additional algorithms in addition to conventional MA. And, in case, you need a particular feature, I could implement it - if time permits.
The software is intended for the meta-analysis of GWASes, hence rather big data. It can handle compressed files (.gz) and can parse (almost) any tabulated result file, by editing a configuration file. Plus, there is an exhaustive command line interface and documentation.
HTH
Christian
edit:
PS Does RevMan really support MA of GWASes? I am asking because this was the question, right?
PPS In case you do not run linux, you could try cygwin on a Windows machine. Anyway, in case of small studies, e.g. replication studies, using R is probably fine, but you would have to code a bit. Here is something to get started: http://www.r-project.org/conferences/useR-2009/slides/Zhao+Tan.pdf
PPPS My software does not offer (good) plotting, as of yet, but I could polish some not yet released features for you. However, for plotting, e.g. Manhattan plots or forest plot (for which I have Python scripts), lots of 3rd party tools are around and it is easy to write custom scripts.
Just realized that my former colleagues only uploaded a binary. This is unfortunate, because I bundled more files (some utility scripts, documentation and, particularly, sample configuration). In case you are interested, just contact me.
If your meta-analysis in the linked topic is the one in question and you have aggregated GWAS data at hand, the phrased question in this other topic is misleading - and therefore the answers. Else, if I misunderstood your question here, my answers are, of course, without value to the topic.
https://www.researchgate.net/post/How_many_studies_must_be_included_in_a_meta-analysis_while_dealing_with_a_cancer_type_of_low_incidence_in_the_world2?_tpcectx=profile_highlights
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis