how to calculate predicted expression using weights from model file

Question

[rheab1230]

Hello,

I am using prediXcan method to generate model based on gtex data for brain frontal tissue. I am able to train the model file and using Predict.py script to generate predicted expression. I want to ask how does the predicted expression is calculated?

I am trying to write my own code to see whether I understood the code part of predixcan.

I am doing it in this way:

gt_Df:
         rs1 rs2    rs8 rs21 
1117F     1    0     1   0 
 IIEI      1   0    1    1
weights file: 

 rs21     rs8         rs2 
0.119   -0.037      0.032

Expected outcome:
              Predicted_Value 
1117F     -0.037 
11EI        0.082 
Calculation: for 1117F: 1*0 + 0*0.032 +1*-0.037 + 0*0.119=-0.037

But the results using prediXcan is different from my calculations. Is there any other way to do the analysis? or I am making any mistake to calculate it.

Thank you.