Hi everyone,
i have been trying to predict risk of developing a cancer from genetic data. At the end I plotted a curve, that shows odds ratio (OR) and relative risk (RR) as a function of polygenic risk score (PRS) - a score computed from genetic data:
At PRS=0 the OR=0.016 and RR=0.026. But for demonstration purposes i would like to translate the curve so the OR and RR at PRS=0 would be 1 (subjects with PRS<0 have OR and RR < 1 and vice versa).
My idea was to simply vertically translate the curve by vector (0, 0.984) for OR and (0, 0.974)for RR. So my question is, if it is statistically correct to vertically translate an OR or/and OR curve by a vector.
Thank you all so much for any help.
Jure
i have been trying to predict risk of developing a cancer from genetic data. At the end I plotted a curve, that shows odds ratio (OR) and relative risk (RR) as a function of polygenic risk score (PRS) - a score computed from genetic data:


At PRS=0 the OR=0.016 and RR=0.026. But for demonstration purposes i would like to translate the curve so the OR and RR at PRS=0 would be 1 (subjects with PRS<0 have OR and RR < 1 and vice versa).
My idea was to simply vertically translate the curve by vector (0, 0.984) for OR and (0, 0.974)for RR. So my question is, if it is statistically correct to vertically translate an OR or/and OR curve by a vector.
Thank you all so much for any help.
Jure
Last edited: