sachalau
commented
1 week ago Dear Woohong,
The overall sensitivity and specificity are calculated by looking back at the per-sample genotype and phenotype results and calculating using the final mutation grades. As we haven't disclosed the per sample results at the moment that is not something you are able to replicate for now, but these should be released soon.
Thank you for sharing this amazing dataset and code. It's really helpful for me to analyze the data.
I am not sure this is the right place to ask this type of question or request, but I wonder how to calculate the combined sensitivity described in the catalogue. For example, in the catalogue, "the combined sensitivity of Groups 1 and 2 mutations in the two catalogues was identical at 93.3% (95% CI: 92.9–93.7) (Table A.1).", but I was not able to reproduce this value. I naively thought the individual sensitivity for Groups 1 and 2 mutations are summed up to get this number, but it was 96.1% which is incorrect. If I do the same analysis for INH, the combined sensitivity I got is >100% which doesn't sound right. Would you please share how I can calculate the correct combined sensitivity? I was also wondering if there is a dataset of allele frequency that I can get access to.
Thank you again!