Date: 2026-04-05 |
Configs: 5 |
Proteins: 10 |
Samples: 26,560 (1,600 finals) |
Runtime: ~93 min on 8x H100 |
PRODUCTION (400 steps)
Key Findings
1. E_fine_selector dominates.
Won 7/10 proteins by best single sample and 7/10 by mean reward.
Its frequent pruning (6 checkpoints, 3 branches) consistently finds the best binders.
2. D_early_brancher drops to 4th with full data (-0.2792).
Its earlier lead was biased by only completing easier proteins. Frequent pruning beats coarse exploration.
3. B_length_explorer is worst (-0.3501 mean reward).
16 lengths with 2 beams each lacks depth. Length diversity doesn't compensate for shallow search.
4. Best binding prediction: IFNAR2 with i_PAE = 0.130 (D_early_brancher).
Hardest target: BetV1 at best i_PAE = 0.205.
5. Lookaheads beat finals in all configs - the search is not optimally selecting
intermediate candidates. Room for improvement in pruning/scoring.
Config x Protein Heatmaps
Mean i_PAE across final samples (lower = better binding). Missing cells = experiment failed.Best single-sample i_PAE. Gold border = winner for that protein target.
Configuration Rankings
Left: mean reward. Center: wins by best sample. Right: wins by mean reward.
Rank
Config
Strategy
Mean Reward
Wins (best)
Wins (mean)
1
E_fine_selector
6 checkpoints, 3 branches
-0.2475
7
7
2
A_balanced
Paper default
-0.2571
0
2
3
C_deep_exploiter
2 lengths, 16 beams
-0.2700
1
1
4
D_early_brancher
2 checkpoints, 8 branches
-0.2792
2
0
5
B_length_explorer
16 lengths, 2 beams
-0.3501
0
0
Per-Protein Comparison
Best single-sample i_PAE by protein target, grouped by config.
Protein Difficulty
Easiest to hardest protein targets. Green < 0.16 (easy), Orange 0.16-0.22 (medium), Red > 0.22 (hard).
Difficulty
Protein
Best i_PAE
Best Config
Easy
IFNAR2
0.130
D_early_brancher
Easy
PD1
0.138
E_fine_selector
Easy
PDL1
0.150
C_deep_exploiter
Easy
CrSAS6
0.156
E_fine_selector
Medium
CD45
0.159
E_fine_selector
Medium
SpCas9
0.168
E_fine_selector
Medium
DerF7
0.175
E_fine_selector
Medium
Claudin1
0.198
C_deep_exploiter
Hard
BetV1
0.205
E_fine_selector
Hard
HER2_AAV
0.222
E_fine_selector
Compute Budget
Total PDBs generated vs finals kept per experiment. E_fine_selector generates the most (608) but wins the most.
Next Steps
Add refinement to E_fine_selector:sequence_hallucination is expected to dramatically improve pLDDT and i_PAE
Fix D_early_brancher failures: Reduce n_branch from 8 to 6 to avoid OOM on larger targets
Full pipeline on top candidates: filter → evaluate → analyze for designability (scRMSD, ProteinMPNN)