Autoresearch Denovo

Trajectory Visualizer — Interactive Dashboard
21
Total Runs
4
Targets
1004
Final Samples
0.1302
Best i_pae
8
Sessions
Target Overview 4 targets
Target Runs Best i_pae
01_PD140.1331
02_PDL180.1344
04_IFNAR250.1302
05_CD4540.1732
Optimization Trajectory
Run Explorer
Insight Timeline
Hyperparameter Map
Agent Timeline
Sample Distributions
Optimization Trajectory best_i_pae over runs
Improvement Rate % from first run
Mean Reward per run
Finals Count n_finals per run
Config Diversity beam_width × n_branch
Analysis Notes multi-persona review 2026-04-07
ML Optimization
The LLM functions as a mediocre hill-climber: same 4 knobs in the same order every target. 3 duplicate PDL1 runs wasted 25% compute. Never tried fk-steering, best-of-n, or nsteps<400. A GP-BO would likely match results in half the runs.
Structural Biology
Best i_pae (0.130-0.134) is marginal vs BindCraft (<0.10 validated). pLDDT is excellent (0.94-0.96). CD45 fails due to 1 hotspot. ProteinMPNN exists but is disabled — highest-impact change.
Systems
GPU utilization ~35-40%. Total: ~16 GPU-hrs ($47). 25% wasted on duplicates. No auto-retry, orphaned process risk, no file locking.
Consensus
(1) Enable ProteinMPNN, (2) GP/TPE surrogate, (3) Fix duplicate-run bug. Full review: memo/2026-04-07_trajectory_review.md
Hyperparameter Exploration Map parallel coordinates
Agent Decision Timeline tool calls per session
Sample i_pae Distributions per-run box plots
Agent Insights & Hypotheses scientific observations
02_PDL1 2026-04-05 20:25
After three beam-search runs with identical beam_width=8, n_branch=4, nsamples=4 and standard checkpoints, best i_pae stalled at ~0.158 and top candidates have low pLDDT (0.04–0.09). Hypothesis: expanding length coverage (increase nsamples) may find better-length binders and improve i_pae; consider increasing nsamples to 8 while keeping other params constant for the next run.
02_PDL1 2026-04-06 02:00
Increasing nsamples from 4 to 8 did not improve best i_pae (worsened from 0.1559 to 0.1714) and pLDDT remains very low (~0.04–0.06), suggesting structural confidence is poor. Hypothesis: the search is stuck in similar low-confidence modes; increasing beam width to 16 to increase sequence/structure diversity may find higher-affinity, higher-confidence binders. Monitor for improved i_pae and pLDDT; if no improvement, consider changing branching (n_branch) or switching search_algorithm.
04_IFNAR2 2026-04-06 02:50
Increasing beam_width from 8 to 16 improved best i_pae from 0.1448 to 0.1387, indicating broader beam search yields better interface predictions. nsamples increase to 8 caused CUDA OOM at current batch size/config, so we should avoid raising nsamples without reducing batch_size or memory usage. Next experimental variable to try: increase branching factor (n_branch) to 8 to increase diversity while keeping beam_width at 16 and nsamples at 4.
01_PD1 2026-04-06 04:37
Fine-grained step_checkpoints ([0,65,130,200,270,340,400]) combined with sampling more binder lengths (nsamples=8) and beam-search (beam_width=8, n_branch=4) produced the best binders for 01_PD1 (best i_pae=0.133). Increasing beam_width or branching (beam_width=16 or n_branch=8) caused CUDA OOM on the available GPU. MCTS did not outperform beam-search. Recommendation: treat the beam-search, fine-checkpoint, nsamples=8 configuration as the baseline for further local refinement; consider increasing nsamples to 16 or exploring small algorithmic tweaks only if more improvement is required.
02_PDL1 2026-04-06 05:17
Best result so far: beam_width=16, n_branch=4, nsamples=4, nsteps=400, fine checkpoints produced best_i_pae=0.13441561 (run_20260406_034234_02_PDL1). However pLDDT remains low (~0.04–0.07). Increasing n_branch to 8 caused CUDA OOM. Budget has reached 10 runs; stop launching new experiments. Recommended next steps: (1) take top candidates from run_20260406_034234_02_PDL1 for structural refinement (AF refinement/relaxation) and scoring; (2) if budget is extended, test search_algorithm=fk-steering or mcts, or try beam_width=16 with n_branch=8 but smaller batch_size to fit memory, aiming to increase diversity and pLDDT.
02_PDL1 2026-04-06 14:13
Best run (run_20260406_034234_02_PDL1) achieved i_pae=0.1344 with beam-search (beam_width=16, n_branch=4, nsamples=4, fine checkpoints), but pLDDT remains very low (~0.04–0.07). Next recommended actions: (1) perform structural refinement (AF-relax/refinement) on top candidates to improve pLDDT and re-score; (2) if running another design, change search algorithm (try fk-steering or mcts) rather than repeating beam-search; avoid increasing n_branch beyond 4 due to OOM risk.
04_IFNAR2 2026-04-06 14:13
Increasing beam_width from 8->16 and n_branch from 4->8 improved best i_pae from 0.1448 to 0.13016. Attempting nsamples=8 previously caused CUDA OOM; keep nsamples=4 unless batch_size is reduced. Next planned change: use fine step_checkpoints [0,65,130,200,270,340,400] to give more branching decision points and refine interface sampling.
04_IFNAR2 2026-04-06 14:41
Recent attempts using fine step_checkpoints caused CUDA OOMs (batch_size=8 and batch_size=4 failed). Memory pressure likely from increased branching checkpoints. Next attempt: keep beam_width=16 and n_branch=8 but reduce batch_size to 2 to fit GPU memory. If this still OOMs, consider reducing beam_width or n_branch or switching to standard checkpoints.
05_CD45 2026-04-06 15:06
For 05_CD45, using beam_width=16 can improve i_pae (best so far 0.173) but requires a small generation batch_size to avoid CUDA OOM. Increasing n_branch to 8 caused OOM with batch_size=8; similarly nsamples=8 with beam_width=16 and batch_size=8 also OOM'd. Recommendation: keep beam_width=16 but reduce batch_size (e.g., 2) or reduce n_branch to 2–3 to fit GPU memory. Also, fine checkpoints improved best scores in earlier runs.
Run Explorer per-run details with LLM reasoning
01_PD1 4 runs best: 0.1331
Run 1 0.1660
bw=8nb=4ns=4beam-searchstandard ckpts (5pt)
2026-04-06T01:38
run_20260406_011745_01_PD1
0.1660 best i_pae-0.1938 mean reward32 finals
Top samples:
#1 i_pae=0.1660 plddt=0.0774 job_0_n_197_id_128_beam_orig3_bm5-s0to100br3-s100to200br3-s2#2 i_pae=0.1669 plddt=0.0784 job_0_n_197_id_129_beam_orig3_bm5-s0to100br3-s100to200br3-s2#3 i_pae=0.1673 plddt=0.0789 job_0_n_197_id_130_beam_orig3_bm5-s0to100br3-s100to200br3-s2#4 i_pae=0.1696 plddt=0.0810 job_0_n_197_id_131_beam_orig3_bm5-s0to100br3-s100to200br3-s2#5 i_pae=0.1696 plddt=0.0811 job_0_n_197_id_132_beam_orig3_bm5-s0to100br3-s100to200br3-s2
Run 2 0.1380
bw=8nb=4ns=4beam-searchfine ckpts (7pt)
2026-04-06T02:08
run_20260406_013858_01_PD1
0.1380 best i_pae-0.1546 mean reward32 finals
Top samples:
#1 i_pae=0.1380 plddt=0.0548 job_0_n_201_id_192_beam_orig3_bm1-s0to65br1-s65to130br0-s130#2 i_pae=0.1383 plddt=0.0548 job_0_n_201_id_194_beam_orig3_bm1-s0to65br1-s65to130br0-s130#3 i_pae=0.1384 plddt=0.0551 job_0_n_201_id_195_beam_orig3_bm1-s0to65br1-s65to130br0-s130#4 i_pae=0.1385 plddt=0.0551 job_0_n_201_id_193_beam_orig3_bm1-s0to65br1-s65to130br0-s130#5 i_pae=0.1416 plddt=0.0563 job_0_n_201_id_196_beam_orig3_bm1-s0to65br1-s65to130br0-s130
Run 3 0.1331
bw=8nb=4ns=8beam-searchfine ckpts (7pt)
2026-04-06T03:12
run_20260406_020918_01_PD1
0.1331 best i_pae-0.1582 mean reward64 finals
Top samples:
#1 i_pae=0.1331 plddt=0.0496 job_0_n_206_id_192_beam_orig6_bm1-s0to65br3-s65to130br0-s130#2 i_pae=0.1331 plddt=0.0496 job_0_n_206_id_193_beam_orig6_bm1-s0to65br3-s65to130br0-s130#3 i_pae=0.1331 plddt=0.0497 job_0_n_206_id_194_beam_orig6_bm1-s0to65br3-s65to130br0-s130#4 i_pae=0.1333 plddt=0.0499 job_0_n_206_id_195_beam_orig6_bm1-s0to65br3-s65to130br0-s130#5 i_pae=0.1376 plddt=0.0512 job_0_n_206_id_198_beam_orig6_bm4-s0to65br2-s65to130br3-s130
Run 4 0.1381
bw=8nb=4ns=8mctsfine ckpts (7pt)
2026-04-06T04:36
run_20260406_035801_01_PD1
0.1381 best i_pae-0.1654 mean reward8 finals
Top samples:
#1 i_pae=0.1381 plddt=0.0472 job_0_n_254_id_120_mcts_orig5-s0to65br4-s65to130br1-s130to20#2 i_pae=0.1498 plddt=0.0376 job_0_n_236_id_120_mcts_orig7-s0to65br3-s65to130br9-s130to20#3 i_pae=0.1567 plddt=0.0644 job_0_n_245_id_120_mcts_orig2-s0to65br8-s65to130br5-s130to20#4 i_pae=0.1662 plddt=0.0465 job_0_n_261_id_120_mcts_orig4-s0to65br5-s65to130br0-s130to20#5 i_pae=0.1753 plddt=0.0607 job_0_n_265_id_120_mcts_orig6-s0to65br0-s65to130br6-s130to20
02_PDL1 8 runs best: 0.1344
Run 1 0.1581
bw=?nb=?ns=?beam-search
2026-04-05T19:15
run_20260405_161907_02_PDL1
0.1581 best i_pae-0.1743 mean reward32 finals
Top samples:
#1 i_pae=0.1581 plddt=0.0465 job_0_n_262_id_128_beam_orig0_bm5-s0to100br2-s100to200br0-s2#2 i_pae=0.1606 plddt=0.0474 job_0_n_262_id_129_beam_orig0_bm5-s0to100br2-s100to200br0-s2#3 i_pae=0.1632 plddt=0.0567 job_0_n_262_id_130_beam_orig0_bm5-s0to100br2-s100to200br0-s2#4 i_pae=0.1651 plddt=0.0585 job_0_n_202_id_128_beam_orig3_bm4-s0to100br1-s100to200br2-s2#5 i_pae=0.1654 plddt=0.0590 job_0_n_202_id_129_beam_orig3_bm4-s0to100br1-s100to200br2-s2
Run 2 0.1581
bw=?nb=?ns=?beam-search
2026-04-05T20:01
run_20260405_191629_02_PDL1
0.1581 best i_pae-0.1743 mean reward32 finals
Top samples:
#1 i_pae=0.1581 plddt=0.0465 job_0_n_262_id_128_beam_orig0_bm5-s0to100br2-s100to200br0-s2#2 i_pae=0.1606 plddt=0.0474 job_0_n_262_id_129_beam_orig0_bm5-s0to100br2-s100to200br0-s2#3 i_pae=0.1632 plddt=0.0567 job_0_n_262_id_130_beam_orig0_bm5-s0to100br2-s100to200br0-s2#4 i_pae=0.1651 plddt=0.0585 job_0_n_202_id_128_beam_orig3_bm4-s0to100br1-s100to200br2-s2#5 i_pae=0.1654 plddt=0.0590 job_0_n_202_id_129_beam_orig3_bm4-s0to100br1-s100to200br2-s2
Run 3 0.1581
bw=?nb=?ns=?beam-search
2026-04-05T20:25
run_20260405_200257_02_PDL1
0.1581 best i_pae-0.1743 mean reward32 finals
Top samples:
#1 i_pae=0.1581 plddt=0.0465 job_0_n_262_id_128_beam_orig0_bm5-s0to100br2-s100to200br0-s2#2 i_pae=0.1606 plddt=0.0474 job_0_n_262_id_129_beam_orig0_bm5-s0to100br2-s100to200br0-s2#3 i_pae=0.1632 plddt=0.0567 job_0_n_262_id_130_beam_orig0_bm5-s0to100br2-s100to200br0-s2#4 i_pae=0.1651 plddt=0.0585 job_0_n_202_id_128_beam_orig3_bm4-s0to100br1-s100to200br2-s2#5 i_pae=0.1654 plddt=0.0590 job_0_n_202_id_129_beam_orig3_bm4-s0to100br1-s100to200br2-s2
Launch2026-04-05T20:02:57
{
  "target": "02_PDL1",
  "beam_width": 8,
  "n_branch": 4,
  "nsamples": 4,
  "nsteps": 400,
  "step_checkpoints": [
    0,
    100,
    200,
    300,
    400
  ],
  "search_algorithm": "beam-search"
}
Results2026-04-05T20:25:34
{
  "best_i_pae": 0.15809426,
  "mean_i_pae": 0.174298794375,
  "best_reward": -0.18657243,
  "mean_reward": -0.174298794375,
  "best_plddt": 0.09077891,
  "mean_plddt": 0.06490734465625,
  "n_total": 544,
  "n_finals": 32,
  "n_lookaheads": 512,
  "top_samples": [
    {
      "pdb_path": "/orcd/pool/006/yeopjin/workspace/Autoresearch_Denovo/subgit/inference/search_binder_local_pipeline_02_PDL1_agent_20260405_200257_02_PDL1/job_0_n_262_id_128_beam_orig0_bm5-s0to100br2-s100to200br0-s200to300br3-s
Run 4 0.1559
bw=?nb=?ns=?beam-search
2026-04-05T21:36
run_20260405_205045_02_PDL1
0.1559 best i_pae-0.1770 mean reward64 finals
Top samples:
#1 i_pae=0.1559 plddt=0.0440 job_0_n_233_id_128_beam_orig2_bm6-s0to100br0-s100to200br3-s2#2 i_pae=0.1564 plddt=0.0433 job_0_n_233_id_130_beam_orig2_bm6-s0to100br0-s100to200br3-s2#3 i_pae=0.1567 plddt=0.0582 job_0_n_234_id_128_beam_orig1_bm0-s0to100br0-s100to200br3-s2#4 i_pae=0.1571 plddt=0.0434 job_0_n_233_id_129_beam_orig2_bm6-s0to100br0-s100to200br3-s2#5 i_pae=0.1610 plddt=0.0587 job_0_n_234_id_129_beam_orig1_bm0-s0to100br0-s100to200br3-s2
Launch2026-04-05T20:50:45
{
  "target": "02_PDL1",
  "beam_width": 8,
  "n_branch": 4,
  "nsamples": 8,
  "nsteps": 400,
  "step_checkpoints": [
    0,
    100,
    200,
    300,
    400
  ],
  "search_algorithm": "beam-search",
  "batch_size": 4
}
Results2026-04-05T21:36:51
{
  "best_i_pae": 0.15587172,
  "mean_i_pae": 0.17695078484375,
  "best_reward": -0.20754825,
  "mean_reward": -0.17695078484375,
  "best_plddt": 0.1089551,
  "mean_plddt": 0.069621715125,
  "n_total": 1088,
  "n_finals": 64,
  "n_lookaheads": 1024,
  "top_samples": [
    {
      "pdb_path": "/orcd/pool/006/yeopjin/workspace/Autoresearch_Denovo/subgit/inference/search_binder_local_pipeline_02_PDL1_agent_20260405_205045_02_PDL1/job_0_n_233_id_128_beam_orig2_bm6-s0to100br0-s100to200br3-s200to300br
LLM Analysis2026-04-05T21:37:09
Summary of completed runs (chronological):
- run_20260405_161907_02_PDL1 — best_i_pae 0.15809426
- run_20260405_191629_02_PDL1 — best_i_pae 0.15809426
- run_20260405_200257_02_PDL1 — best_i_pae 0.15809426 (nsamples=4)
- run_20260405_202553_02_PDL1 — FAILED (nsamples=8, batch_size=8) due to CUDA OOM
- run_20260405_205045_02_PDL1 — COMPLETED: best_i_pae = 0.15587172, best_plddt = 0.1089551 (nsamples=8, batch_size=4)

Key observations
- Increasing nsamples from 4 → 8 gave a measurable improvement in best i_pae (0.1581 → 0.1559) and a small increase in best pLDDT.
- Top candidate pLDDT values rema
Run 5 0.1714
bw=8nb=4ns=8beam-searchstandard ckpts (5pt)
2026-04-06T02:00
run_20260406_011750_02_PDL1
0.1714 best i_pae-0.2289 mean reward64 finals
Top samples:
#1 i_pae=0.1714 plddt=0.0456 job_0_n_225_id_128_beam_orig2_bm0-s0to100br2-s100to200br1-s2#2 i_pae=0.1759 plddt=0.0473 job_0_n_225_id_129_beam_orig2_bm0-s0to100br2-s100to200br1-s2#3 i_pae=0.1760 plddt=0.0464 job_0_n_225_id_130_beam_orig2_bm0-s0to100br2-s100to200br1-s2#4 i_pae=0.1764 plddt=0.0594 job_0_n_240_id_128_beam_orig4_bm1-s0to100br2-s100to200br0-s2#5 i_pae=0.1779 plddt=0.0606 job_0_n_240_id_129_beam_orig4_bm1-s0to100br2-s100to200br2-s2
Launch2026-04-06T01:17:50
{
  "target": "02_PDL1",
  "beam_width": 8,
  "n_branch": 4,
  "nsamples": 8,
  "nsteps": 400,
  "step_checkpoints": [
    0,
    100,
    200,
    300,
    400
  ],
  "search_algorithm": "beam-search"
}
Results2026-04-06T02:00:35
{
  "best_i_pae": 0.17136922,
  "mean_i_pae": 0.22892677078125,
  "best_reward": -0.48377234,
  "mean_reward": -0.22892677078125,
  "best_plddt": 0.19808209,
  "mean_plddt": 0.087950295328125,
  "n_total": 1088,
  "n_finals": 64,
  "n_lookaheads": 1024,
  "top_samples": [
    {
      "pdb_path": "/orcd/pool/006/yeopjin/workspace/Autoresearch_Denovo/subgit/inference/search_binder_local_pipeline_02_PDL1_agent_20260406_011750_02_PDL1/job_0_n_225_id_128_beam_orig2_bm0-s0to100br2-s100to200br1-s200to3
Run 6 0.1648
bw=16nb=4ns=4beam-searchstandard ckpts (5pt)
2026-04-06T03:11
run_20260406_022337_02_PDL1
0.1648 best i_pae-0.1788 mean reward64 finals
Top samples:
#1 i_pae=0.1648 plddt=0.0612 job_0_n_243_id_256_beam_orig1_bm12-s0to100br3-s100to200br1-s#2 i_pae=0.1660 plddt=0.0652 job_0_n_245_id_256_beam_orig2_bm13-s0to100br0-s100to200br2-s#3 i_pae=0.1661 plddt=0.0620 job_0_n_243_id_257_beam_orig1_bm12-s0to100br3-s100to200br1-s#4 i_pae=0.1670 plddt=0.0644 job_0_n_243_id_259_beam_orig1_bm12-s0to100br3-s100to200br1-s#5 i_pae=0.1671 plddt=0.0604 job_0_n_198_id_257_beam_orig3_bm9-s0to100br3-s100to200br3-s2
Launch2026-04-06T02:23:37
{
  "target": "02_PDL1",
  "beam_width": 16,
  "n_branch": 4,
  "nsamples": 4,
  "nsteps": 400,
  "step_checkpoints": [
    0,
    100,
    200,
    300,
    400
  ],
  "search_algorithm": "beam-search",
  "batch_size": 4
}
Run 7 0.1497
bw=8nb=4ns=4beam-searchfine ckpts (7pt)
2026-04-06T03:41
run_20260406_031124_02_PDL1
0.1497 best i_pae-0.1583 mean reward32 finals
Top samples:
#1 i_pae=0.1497 plddt=0.0545 job_0_n_234_id_193_beam_orig0_bm2-s0to65br1-s65to130br2-s130#2 i_pae=0.1498 plddt=0.0545 job_0_n_234_id_192_beam_orig0_bm2-s0to65br1-s65to130br2-s130#3 i_pae=0.1499 plddt=0.0545 job_0_n_234_id_195_beam_orig0_bm2-s0to65br1-s65to130br2-s130#4 i_pae=0.1501 plddt=0.0544 job_0_n_234_id_194_beam_orig0_bm2-s0to65br1-s65to130br2-s130#5 i_pae=0.1518 plddt=0.0596 job_0_n_234_id_197_beam_orig0_bm2-s0to65br1-s65to130br0-s130
Launch2026-04-06T03:11:24
{
  "target": "02_PDL1",
  "beam_width": 8,
  "n_branch": 4,
  "nsamples": 4,
  "nsteps": 400,
  "step_checkpoints": [
    0,
    65,
    130,
    200,
    270,
    340,
    400
  ],
  "search_algorithm": "beam-search"
}
Results2026-04-06T03:42:09
{
  "best_i_pae": 0.14973082,
  "mean_i_pae": 0.1582596615625,
  "best_reward": -0.16846307,
  "mean_reward": -0.1582596615625,
  "best_plddt": 0.07117624,
  "mean_plddt": 0.053620649,
  "n_total": 800,
  "n_finals": 32,
  "n_lookaheads": 768,
  "top_samples": [
    {
      "pdb_path": "/orcd/pool/006/yeopjin/workspace/Autoresearch_Denovo/subgit/inference/search_binder_local_pipeline_02_PDL1_agent_20260406_031124_02_PDL1/job_0_n_234_id_193_beam_orig0_bm2-s0to65br1-s65to130br2-s130to200br3-s200to
Run 8 0.1344
bw=16nb=4ns=4beam-searchfine ckpts (7pt)
2026-04-06T04:50
run_20260406_034234_02_PDL1
0.1344 best i_pae-0.1564 mean reward64 finals
Top samples:
#1 i_pae=0.1344 plddt=0.0417 job_0_n_241_id_384_beam_orig2_bm12-s0to65br3-s65to130br3-s13#2 i_pae=0.1345 plddt=0.0417 job_0_n_241_id_385_beam_orig2_bm12-s0to65br3-s65to130br3-s13#3 i_pae=0.1345 plddt=0.0417 job_0_n_241_id_387_beam_orig2_bm12-s0to65br3-s65to130br3-s13#4 i_pae=0.1346 plddt=0.0418 job_0_n_241_id_386_beam_orig2_bm12-s0to65br3-s65to130br3-s13#5 i_pae=0.1397 plddt=0.0468 job_0_n_241_id_389_beam_orig2_bm12-s0to65br3-s65to130br3-s13
Context2026-04-06T03:42:09
Prior run result: best_i_pae=0.14973082
Launch2026-04-06T03:42:34
{
  "target": "02_PDL1",
  "beam_width": 16,
  "n_branch": 4,
  "nsamples": 4,
  "nsteps": 400,
  "step_checkpoints": [
    0,
    65,
    130,
    200,
    270,
    340,
    400
  ],
  "search_algorithm": "beam-search",
  "batch_size": 4
}
Results2026-04-06T04:50:20
{
  "best_i_pae": 0.13441561,
  "mean_i_pae": 0.15642489375,
  "best_reward": -0.1666406,
  "mean_reward": -0.15642489375,
  "best_plddt": 0.06493132,
  "mean_plddt": 0.05324510603125,
  "n_total": 1600,
  "n_finals": 64,
  "n_lookaheads": 1536,
  "top_samples": [
    {
      "pdb_path": "/orcd/pool/006/yeopjin/workspace/Autoresearch_Denovo/subgit/inference/search_binder_local_pipeline_02_PDL1_agent_20260406_034234_02_PDL1/job_0_n_241_id_384_beam_orig2_bm12-s0to65br3-s65to130br3-s130to200br2-s20
04_IFNAR2 5 runs best: 0.1302
Run 1 0.1448
bw=8nb=4ns=4beam-searchstandard ckpts (5pt)
2026-04-06T01:43
run_20260406_011745_04_IFNAR2
0.1448 best i_pae-0.1524 mean reward32 finals
Top samples:
#1 i_pae=0.1448 plddt=0.0470 job_0_n_251_id_128_beam_orig1_bm0-s0to100br0-s100to200br0-s2#2 i_pae=0.1449 plddt=0.0473 job_0_n_251_id_129_beam_orig1_bm0-s0to100br0-s100to200br0-s2#3 i_pae=0.1459 plddt=0.0477 job_0_n_251_id_130_beam_orig1_bm0-s0to100br0-s100to200br0-s2#4 i_pae=0.1469 plddt=0.0486 job_0_n_251_id_131_beam_orig1_bm0-s0to100br0-s100to200br1-s2#5 i_pae=0.1480 plddt=0.0497 job_0_n_251_id_132_beam_orig1_bm0-s0to100br0-s100to200br1-s2
Run 2 0.1387
bw=16nb=4ns=4beam-searchstandard ckpts (5pt)
2026-04-06T02:23
run_20260406_014407_04_IFNAR2
0.1387 best i_pae-0.1497 mean reward64 finals
Top samples:
#1 i_pae=0.1387 plddt=0.0406 job_0_n_232_id_256_beam_orig3_bm4-s0to100br2-s100to200br2-s2#2 i_pae=0.1391 plddt=0.0425 job_0_n_232_id_259_beam_orig3_bm4-s0to100br2-s100to200br2-s2#3 i_pae=0.1393 plddt=0.0411 job_0_n_232_id_257_beam_orig3_bm4-s0to100br2-s100to200br2-s2#4 i_pae=0.1394 plddt=0.0411 job_0_n_232_id_258_beam_orig3_bm4-s0to100br2-s100to200br2-s2#5 i_pae=0.1429 plddt=0.0407 job_0_n_232_id_260_beam_orig3_bm4-s0to100br2-s100to200br2-s2
Run 3 0.1302
bw=16nb=8ns=4beam-searchstandard ckpts (5pt)
2026-04-06T05:38
run_20260406_035723_04_IFNAR2
0.1302 best i_pae-0.1386 mean reward64 finals
Top samples:
#1 i_pae=0.1302 plddt=0.0406 job_0_n_199_id_512_beam_orig1_bm0-s0to100br7-s100to200br5-s2#2 i_pae=0.1313 plddt=0.0447 job_0_n_199_id_513_beam_orig1_bm0-s0to100br7-s100to200br5-s2#3 i_pae=0.1326 plddt=0.0465 job_0_n_199_id_514_beam_orig1_bm0-s0to100br7-s100to200br5-s2#4 i_pae=0.1327 plddt=0.0454 job_0_n_199_id_515_beam_orig1_bm0-s0to100br7-s100to200br5-s2#5 i_pae=0.1330 plddt=0.0426 job_0_n_199_id_516_beam_orig1_bm0-s0to100br7-s100to200br5-s2
Run 4 0.1459
bw=16nb=8ns=4mctsstandard ckpts (5pt)
2026-04-06T16:40
run_20260406_162403_04_IFNAR2
0.1459 best i_pae-0.1734 mean reward4 finals
Top samples:
#1 i_pae=0.1459 plddt=0.0607 job_0_n_179_id_80_mcts_orig1-s0to100br1-s100to200br1-s200to3#2 i_pae=0.1624 plddt=0.1063 job_0_n_167_id_80_mcts_orig3-s0to100br7-s100to200br10-s200to#3 i_pae=0.1840 plddt=0.0851 job_0_n_277_id_80_mcts_orig2-s0to100br1-s100to200br8-s200to3#4 i_pae=0.2011 plddt=0.0735 job_0_n_235_id_80_mcts_orig0-s0to100br4-s100to200br4-s200to3
Launch2026-04-06T16:24:03
{
  "target": "04_IFNAR2",
  "beam_width": 16,
  "n_branch": 8,
  "nsamples": 4,
  "nsteps": 400,
  "step_checkpoints": [
    0,
    100,
    200,
    300,
    400
  ],
  "search_algorithm": "mcts"
}
Results2026-04-06T16:40:57
{
  "best_i_pae": 0.14587466,
  "mean_i_pae": 0.173355985,
  "best_reward": -0.20114449,
  "mean_reward": -0.173355985,
  "best_plddt": 0.10626192,
  "mean_plddt": 0.08140026575,
  "n_total": 324,
  "n_finals": 4,
  "n_lookaheads": 320,
  "top_samples": [
    {
      "pdb_path": "/orcd/pool/006/yeopjin/workspace/Autoresearch_Denovo/subgit/inference/search_binder_local_pipeline_04_IFNAR2_agent_20260406_162403_04_IFNAR2/job_0_n_179_id_80_mcts_orig1-s0to100br1-s100to200br1-s200to300br9-s300to400br2
Run 5 0.1332
bw=16nb=8ns=8beam-searchstandard ckpts (5pt)
2026-04-07T01:29
run_20260406_180022_04_IFNAR2
0.1332 best i_pae-0.1439 mean reward128 finals
Top samples:
#1 i_pae=0.1332 plddt=0.0573 job_0_n_175_id_514_beam_orig0_bm6-s0to100br4-s100to200br6-s2#2 i_pae=0.1333 plddt=0.0577 job_0_n_175_id_515_beam_orig0_bm6-s0to100br4-s100to200br6-s2#3 i_pae=0.1333 plddt=0.0574 job_0_n_175_id_512_beam_orig0_bm6-s0to100br4-s100to200br6-s2#4 i_pae=0.1333 plddt=0.0565 job_0_n_175_id_516_beam_orig0_bm6-s0to100br4-s100to200br6-s2#5 i_pae=0.1335 plddt=0.0574 job_0_n_175_id_513_beam_orig0_bm6-s0to100br4-s100to200br6-s2
Launch2026-04-06T18:00:22
{
  "target": "04_IFNAR2",
  "beam_width": 16,
  "n_branch": 8,
  "nsamples": 8,
  "nsteps": 400,
  "step_checkpoints": [
    0,
    100,
    200,
    300,
    400
  ],
  "search_algorithm": "beam-search",
  "batch_size": 1
}
05_CD45 4 runs best: 0.1732
Run 1 0.2412
bw=8nb=4ns=4beam-searchstandard ckpts (5pt)
2026-04-06T01:41
run_20260406_011746_05_CD45
0.2412 best i_pae-0.4120 mean reward32 finals
Top samples:
#1 i_pae=0.2412 plddt=0.0604 job_0_n_255_id_128_beam_orig1_bm2-s0to100br0-s100to200br0-s2#2 i_pae=0.2438 plddt=0.0644 job_0_n_255_id_130_beam_orig1_bm2-s0to100br0-s100to200br0-s2#3 i_pae=0.2457 plddt=0.0635 job_0_n_255_id_129_beam_orig1_bm2-s0to100br0-s100to200br0-s2#4 i_pae=0.2551 plddt=0.0625 job_0_n_255_id_131_beam_orig1_bm2-s0to100br0-s100to200br0-s2#5 i_pae=0.2607 plddt=0.0690 job_0_n_261_id_129_beam_orig3_bm1-s0to100br3-s100to200br1-s2
Run 2 0.1796
bw=8nb=4ns=4beam-searchfine ckpts (7pt)
2026-04-06T02:53
run_20260406_022400_05_CD45
0.1796 best i_pae-0.2042 mean reward32 finals
Top samples:
#1 i_pae=0.1796 plddt=0.0432 job_0_n_220_id_195_beam_orig2_bm2-s0to65br0-s65to130br0-s130#2 i_pae=0.1802 plddt=0.0434 job_0_n_220_id_192_beam_orig2_bm2-s0to65br0-s65to130br0-s130#3 i_pae=0.1804 plddt=0.0433 job_0_n_220_id_194_beam_orig2_bm2-s0to65br0-s65to130br0-s130#4 i_pae=0.1808 plddt=0.0435 job_0_n_220_id_193_beam_orig2_bm2-s0to65br0-s65to130br0-s130#5 i_pae=0.1823 plddt=0.0416 job_0_n_220_id_198_beam_orig2_bm2-s0to65br0-s65to130br0-s130
Run 3 0.1805
bw=8nb=4ns=8beam-searchfine ckpts (7pt)
2026-04-06T04:08
run_20260406_025359_05_CD45
0.1805 best i_pae-0.2467 mean reward64 finals
Top samples:
#1 i_pae=0.1805 plddt=0.0503 job_0_n_229_id_193_beam_orig1_bm3-s0to65br3-s65to130br1-s130#2 i_pae=0.1808 plddt=0.0504 job_0_n_229_id_192_beam_orig1_bm3-s0to65br3-s65to130br1-s130#3 i_pae=0.1817 plddt=0.0509 job_0_n_229_id_195_beam_orig1_bm3-s0to65br3-s65to130br1-s130#4 i_pae=0.1829 plddt=0.0511 job_0_n_229_id_194_beam_orig1_bm3-s0to65br3-s65to130br1-s130#5 i_pae=0.1852 plddt=0.0455 job_0_n_250_id_193_beam_orig1_bm5-s0to65br0-s65to130br2-s130
Run 4 0.1732
bw=16nb=4ns=4beam-searchfine ckpts (7pt)
2026-04-06T05:57
run_20260406_042541_05_CD45
0.1732 best i_pae-0.3222 mean reward64 finals
Top samples:
#1 i_pae=0.1732 plddt=0.0373 job_0_n_237_id_386_beam_orig0_bm4-s0to65br2-s65to130br3-s130#2 i_pae=0.1740 plddt=0.0372 job_0_n_237_id_384_beam_orig0_bm4-s0to65br2-s65to130br3-s130#3 i_pae=0.1746 plddt=0.0373 job_0_n_237_id_387_beam_orig0_bm4-s0to65br2-s65to130br3-s130#4 i_pae=0.1751 plddt=0.0375 job_0_n_237_id_385_beam_orig0_bm4-s0to65br2-s65to130br3-s130#5 i_pae=0.1798 plddt=0.0626 job_0_n_237_id_388_beam_orig0_bm4-s0to65br2-s65to130br3-s130