This paper provides an analysis of trade-off of purity and yield for diffusive membrane separations - high performance counter-current membrane purification (HCPMP). This is a great paper for bioprocessing and can be applied to new modalities (cell and gene therapy) as well as traditional mAbs. I think it could be particularly useful for ADC processing too - to separate the mAb from ADC. The model development is presented clearly and intuitively. I recommend publication of this paper with some suggested minor revisions:

1. Most of the data is presented for a selectivity of 70. What is the minimum selectivity needed to achieve meaningful separation for a system of say, Myoglobin and BSA? What is the value of selectivity for which the value doesn't really matter (ie, is selectivity of say 40 different than 70, or does it matter if it is 70 or 7000?). I can intuitively guess this from the equations, but it will be good to call it out.

2. In the corresponding author's previous publication in 2005 - Permeability and selectivity analysis in membranes, which I believe is a pivotal paper in membrane separations - the authors talk about the selectivity permeability trade-offs in membranes in general. Does this technology allow for breaking that rule of thumb for selective separation? I think it might be good to comment on it.

3. How would this technology work for concentrated protein solutions (>10 g/L)? I can see that since there is no net UF, concentration polarization effects are negligible, but still, since we are talking about hindered transport and diffusion coefficients, that do depend on concentration, I would recommend discussing it briefly.

4. I like the comparison between HCPMP and HPTFF. For a long time, many industry professionals including myself thought that HPTFF is a break-through, but actual implementation proved challenging because of HPTFF needing charged membranes (which is not an option in bioprocessing). So:
   a) If we are primarily using a size based separation, what is the MW difference needed for effective separation? 10-fold? That was the rule of thumb in the 1990s for using membranes for selective separation. Even though HCPMP is different than HPTFF, the sieving through the membrane for HCPMP is actually the instrinsic (actual) sieving coefficient. Correct? So a significant size difference should exist. Where I am going is, can we use this technology to separate like sized HCPs from mAbs?
   b) Curious how this HCPMP technology would work if there was a charged membrane.

5. Can the authors comment on using this technology for inline, single pass diafiltration? Seems like that can work.

The authors have addressed all my concerns. The revised version is stronger and will make a great contribution to literature.