Modeling "unpaired" spatial data


#1

Hi, I have a very “easy” problem that I cannot put into brms. To simplyfy: I have expression data for 2 proteins along 20 consecutive cells, I would like to model a correlation (linear, non-linear, ode…) between these 2 proteins. However, measurement for these 2 protein for the same cell is not paired, meaning that protein 1 has been measured for a number of replicates, while protein 2 has been measured seprately for a different number of replicates.
Obviously, I cannot do something like brms(protein1 ~ s(cell) + protein2), as the data for the two proteins is not connected, I have for example 10 detections for protein1 and 6 detection for protein2 in the same cell.
How can I tackle this problem?

thank you


#2

Perhaps the new subset feature (available in the github version of brms) can help you solving this problem. See https://github.com/paul-buerkner/brms/issues/360