Post by jrm23 on May 7, 2019 15:53:47 GMT
Hello,
Thank you to the developers for creating and maintaining such excellent software.
I have question on the COJO analysis module. In the instructions, there is a note that "For a case-control study, the effect size should be log(odds ratio) with its corresponding standard error."
When one runs a COJO analysis (--cojo-slct, --cojo-joint, or --cojo-cond) passing in log(odds ratio) and the corresponding standard error as "b" and "se" in the input.ma file, do the "b*" values in the output (bJ or bC depending on the analysis) correspond to log odds as well? Or, do these "b*" values need to be converted to log(odds ratio) for instance as described in the following links (https://data.broadinstitute.org/alkesgroup/BOLT-LMM/#x1-5200010.2 - section 10.2, or cnsgenomics.com/shiny/LMOR/).
Thank you!
Thank you to the developers for creating and maintaining such excellent software.
I have question on the COJO analysis module. In the instructions, there is a note that "For a case-control study, the effect size should be log(odds ratio) with its corresponding standard error."
When one runs a COJO analysis (--cojo-slct, --cojo-joint, or --cojo-cond) passing in log(odds ratio) and the corresponding standard error as "b" and "se" in the input.ma file, do the "b*" values in the output (bJ or bC depending on the analysis) correspond to log odds as well? Or, do these "b*" values need to be converted to log(odds ratio) for instance as described in the following links (https://data.broadinstitute.org/alkesgroup/BOLT-LMM/#x1-5200010.2 - section 10.2, or cnsgenomics.com/shiny/LMOR/).
Thank you!