The difference between a PSAM and PWM

The difference between a PWM and a PSAM is that a PWM models occupancy and the probability of binding while a PSAM models affinity. Therefore, the PWM assumes that there is a linear relationship between the occupancy of a protein to a binding site and its inherent affinity for the binding site. This linear relationship mostly holds true when the concentration of sequence-specific binding sites far exceeds the free (unbound) protein concentration – which is referred to as being in the “low free protein concentration regime“.

However, when the occupancy of the highest-affinity binding sites starts to saturate then the occupancy-affinity linear relationship between the highest-affinity binding sites and medium affinity sites breaks down. Then, it’s necessary to include the non-negligible free protein concentration in the nonlinear calculation of binding site occupancy based upon both the protein concentration and the protein-DNA affinity.

Multiple studies have shown that very specific proteins do exhibit saturation of their highest affinity binding sites in in vivo cellular conditions. Therefore, being able to model binding saturation with an affinity model like a PSAM can produce more accurate binding estimates.