Meyer-DH / AgingClock

MIT License
16 stars 6 forks source link

Published Results Clarification #3

Open emyr666 opened 1 month ago

emyr666 commented 1 month ago

I am modifying the program to use the polars data frame and to have it seamlessly handle input files where the data uses gene symbols instead of ids, or even a mix. I want to verify that I get the same results as your program for the data you provided : GSE65765_CPM.csv

When I run your biological_age_prediction.py on this input file it does not give the values that you provide in bio_age_example.csv

I.e the published results file is

Run,Bio_Age SRR1793991,26.5714285714 SRR1793992,26.5714285714 SRR1793993,212.5714285714 SRR1793994,318.8571428571

But the results of your program on the data you provide is

        Predicted_Biological_Age

Sample SRR1793991 14.950266 SRR1793992 32.134522 SRR1793993 182.353657 SRR1793994 251.269911

These numbers are quite different. Can you publish an input data file, the expected output on that data from the predictor and then also the expected output from the corrector when using your published predicted result.

emyr666 commented 1 month ago

Just to confirm, for GSE65765_CPM.csv you are using the following values

intercept : 103.54631743289005 median age : 372 median age standard deviation : 64

I see that in issue 1 you say the intercept for humans should be 59.626. Whar should be the values for the median age and it's standard deviation. For the worms you are using units of hours and for humans you are using units of years correct ?

Meyer-DH commented 1 month ago

Hi, your predictions are correct. Sorry, the uploaded file is a bit confusing, the "Bio_Age" is the biological age as computed by the median lifespan and the second correction method. I will update this soon and also upload an updated clock that is more robust.