search

auerbachs / BMDExpress-3

1 stars 0 forks source link

curve fit prefilter and benchmark dose analysis #2

Open wangjiawen2013 opened 11 months ago

wangjiawen2013 commented 11 months ago

Hi, I have tried the latest BMDExpress3 and find the BMD already existed after performing "curve fit prefilter": image then I continuted to perform "benchmark dose analysis", a new BMD was generated: image

What's the difference between these two values ? Which one is the real BMD ?

auerbachs commented 11 months ago

The curve fit prefilter is not intended to derive a BMD. It is meant as means to filtering the probes/genes to identify those that exhibit dose responsive behavior. It is similar to in nature to the other prefilters. We report the BMD on in the Curve fit prefilter results to show that the BMD was within the dose range of the study. The BMD analysis results (e.g. ToxicRMA) is where you want to derive your BMD from (i.e., the best BMD column in the BMD Analysis results).

wangjiawen2013 commented 11 months ago

Hi, I am still confused and want to know more details. The "BEST BMD" in the top figure differs a lot with "HILL BMD" in the bottom figure, which one should I select as the final BMD ? Why there are lots of NaN in BMD analysis results ? And why they are all "HILL" in the BMD analysis results ? Perhaps these are questions about the principle of how ToxicR works. Hope you can give some guidance or references !