TABLE 2.
Dataset | Anti-lysozyme mAbs | CoV-AbDab (test) | EV | NVA1 | NVA2 |
---|---|---|---|---|---|
Antibodies in set | 53 | 7,685 | 126 | 31 | 33 |
Fraction of antibodies modeled | 1.0 | 1.0 | 0.87 | 1.0 | 1.0 |
Fraction of consistent clusters | 1.0 | 0.85 | 0.78 | 0.83 | 1.0 |
Fraction of clustered antibodies in consistent clusters | 1.0 | 0.80 | 0.74 | 0.86 | 1.0 |
Multiple-occupancy clusters | 5 | 1,267 | 9 | 6 | 5 |
Antibodies in multiple-occupancy clusters | 50 (94%) | 4,188 (54%) | 19 (15%) | 14 (45%) | 16 (48%) |
Antibodies in consistent multiple-occupancy clusters | 50 (94%) | 3,353 (44%) | 14 (11%) | 12 (39%) | 16 (48%) |
Values of the five performance metrics and the fraction of antibodies successfully modeled using ABodyBuilder2 are shown for each dataset. For the two metrics of number of antibodies in multiple-occupancy clusters and number of antibodies in epitope-consistent multiple-occupancy clusters, a percentage is shown additionally, indicating the percentage of antibodies in the dataset. CoV-AbDab (test) denotes the subset of CoV-AbDab that is not contained in the Cao et al. (2023) training set (see Materials and Methods). CoV-AbDab (test) was used for this analysis to prevent testing on training set antibodies.