search

ictr / mendel_penetrance

0 stars 1 forks source link

How to estimate confidence intervals and perform single-outcome analysis #5

Open BoPeng opened 8 months ago

BoPeng commented 8 months ago

Hi, @XinYang6699

We have been using this modified mendel program to estimate the over all RR for LFS_related cancers (multiple outcomes), and have the following questions

  1. Could you confirm if this method estimate the RR to get the first cancer, and ignores the rest of the cancers?
  2. What is the role for "other cancer" in the model? We use "other cancer" for non-LFS related cancers and would like to confirm that it contribute to the "base line risk", not the risk for getting LFS-related cancers.
  3. How do we estimate the confidence interval for the estimated risks?
  4. We would like to estimate RR for a single cancer (e.g. lung cancer), how should we prepare the data and interpret the results. My understanding is that we need to set age of onset to all cancers other than the cancer of interest to zero, but not sure if this is the righ approach. In particular, a large number of pedigrees will have all zero age of onset because they do not have any incidences of lung cancer.

Thanks, Bo