Thanks for your outstanding work to collecting these datasets!
I have a problem regarding the direction of favorable values in binding dataset. For example, "AbRank: A Benchmark Dataset and Metric-Learning Framework for Antibody-Antigen Affinity Ranking" has published 342357 antibody-antigen pairs with the binding values measured in "Kd [nM], IC50 [ug/mL], escape". However, according to the paper, the smaller the values, the better the binding ability. Thus, I believe the direction of favorable values regarding this work should be ↓, but in the table you have provided the direction is ↑. I wonder if there is something wrong with this table?
Similar problems are still exist in other lines in the table of binding, e.g.
Measuring the sequence-affinity landscape of antibodies with massively parallel titration curves
Retrospective SARS-CoV-2 human antibody development trajectories are largely sparse and permissive
An integrated technology for quantitative wide mutational scanning of human antibody Fab libraries...
Could you help me take a check on these directions? Thanks again for your great contribution!
Thanks for your outstanding work to collecting these datasets!
I have a problem regarding the direction of favorable values in binding dataset. For example, "AbRank: A Benchmark Dataset and Metric-Learning Framework for Antibody-Antigen Affinity Ranking" has published 342357 antibody-antigen pairs with the binding values measured in "Kd [nM], IC50 [ug/mL], escape". However, according to the paper, the smaller the values, the better the binding ability. Thus, I believe the direction of favorable values regarding this work should be ↓, but in the table you have provided the direction is ↑. I wonder if there is something wrong with this table?
Similar problems are still exist in other lines in the table of binding, e.g.
Measuring the sequence-affinity landscape of antibodies with massively parallel titration curves
Retrospective SARS-CoV-2 human antibody development trajectories are largely sparse and permissive
An integrated technology for quantitative wide mutational scanning of human antibody Fab libraries...
Could you help me take a check on these directions? Thanks again for your great contribution!