Dear all,
I am interested in calculating fractional concentration for a certain protein isoform. For example, I have expression intensity value with log2 transformation for isoform A, B and C. The amount of these isoform is represent as θa, θb and θc.
Given the expression signal S and true probe abundance θ, we know that
log(S)=log(θ)+φ+ε
where φ is a probe-specific effect, and ε accounts for measurement error.
If I compute fractional concentration according to the equation below
fractional concentration=θi/(θ1+θ2+⋯+θn )=Si/(S1+S2+⋯+Sn )
where n is the number of isoform, here I assume that φ+ε=0
I could know that the amount of signal intensity greatly affects the fractional concentration for one and two channel platform, because for one channel their intensity value after normalization and log2 transformation is around 0 and for two channel their intensity value is usually around 7.
So in this condition, how could I transform the data for analysis? I am thinking z-score transformation of the data, if this is scientifically meaningful?
In addition, I wonder if I need to do log2 transformation of the original signal intensity data in calculating the fractional concentration?
Sorry for a long question and I am looking forward your reply!
Thanks a lot!
Best wishes,
Zhe Liu
