Extended Data Fig. 2: Fine-tuned RoseTTAFold2 can distinguish true complexes from decoy complexes.
From: Atomically accurate de novo design of antibodies with RFdiffusion
A-B) Cherry-picked example of RoseTTAFold2 correctly distinguishing a “true” from a “decoy” complex. The sequence of antibody 7Y1B was provided either with the correct (PDB ID: 7Y1B) or decoy (PDB ID: 8CAF) target. Both with 100% (A) or 10% (B) of “hotspots” provided, RF2 near-perfectly predicts binding (top row) or non-binding (bottom row). C) Fine-tuned RoseTTAFold2 reliably predicts its own accuracy. Correlation between RF2 predicted aligned error (pAE) and RMSD to the native structure with 100% (top) or 10% (bottom) of “hotspot” residues provided. With mean pAE <10, 80.3% of structures are within 2 Å when 100% of “hotspots” are provided (along with the holo target structure), with this falling to 52.6% when only 10% of hotspots are provided. D) Quantification of the fine-tuned RF2’s ability to distinguish true targets from decoy targets with both pAE (top row) and pBind (bottom row). Note that this ability depends on the proportion of “hotspots” provided. Without any “hotspots” provided, RF2 is hardly predictive, because RF2 without privileged information is quite rarely confident or accurate in its antibody complex predictions. E-F) Fine-tuned RoseTTAFold is also performant at antibody monomer prediction. 86 antibodies released after the RF2, AlphaFold and IgFold training dataset date cutoff (January 13th, 2023) that share <30% target sequence similarity to any antibody complex released prior to this date were predicted as monomers with either fine-tuned RF2, IgFold, AF2 and AF3. E) The median Fv quality prediction (by overall RMSD) of fine-tuned RF2, of (PDB ID: 8GPG), with (right) and without (left) sidechains shown (gray: native; colors: prediction). While the backbone RMSD is close to the true structure, some sidechains are incorrectly positioned. F) Fine-tuned RF2 slightly outperforms IgFold at prediction accuracy. Overall prediction accuracy is slightly improved in fine-tuned RF2 vs IgFold (p = 0.026, Student’s Paired T-test), with greater improvements in CDR H3 prediction accuracy (p = 0.0003, Student’s Paired T-test).
