In an open microfluidic device, we investigate protein polyacrylamide gel electrophoresis (PAGE) separation performance on single-cell lysate. Single-cell protein electrophoresis is performed in a thin layer of polyacrylamide (PA) gel into which microwells are molded. Individual cells are isolated in these open microwells, then lysed on-chip with a dual lysis and electrophoresis sodium dodecyl sulfate (SDS) buffer. We scrutinize the effect of sieving gel composition on electromigration of protein targets, using a wide range of cellular protein standards (36 kDa to 289 kDa). We find that as PA concentration increases, protein electromigration deviates from the empirical log-linear relationship predicted between migration distance and molecular mass. We perform Ferguson analysis to calculate retardation coefficients and free solution mobilities of nine cellular protein standards and observe that the largest-molecular-mass protein, mTOR (289 kDa), does not behave as predicted by established linear-fit models for SDS-denatured proteins, indicating that mTOR is beyond the linear range of this assay. Lastly, we performed in-gel immunoprobing on the single-cell electrophoretic separations and observed that smaller pore-size gels (higher gel concentration) reduce protein diffusion out of the gel, which does not notably impact the measured immunoprobed protein expression. Compared to larger pore-size gels, the smaller pore-size gels lead to higher local concentrations of the target protein in each protein band, resulting in an increase in the signal-to-noise ratio (SNR) for each protein. Understanding the separation and immunoprobing performance at different gel concentrations improves assay design and optimization for target proteins.