Hypothesis #1359

Bevacizumab VEGF Alzheimer's disease

Drug domain: Glioblastoma Outcome domain: Alzheimer Score: 0.289 Quality: high

← Back to all hypotheses

Provenance trace

Audit trail · Bevacizumab → VEGF → Alzheimer's disease

7 source papers, 7 extracted claims, scored by Swanson ABC chain.

Backfilled

Backfilled retroactively from DB state on 2026-05-13. The pipeline_run this row references was created by the backfill script, not by the actual generation orchestrator. Source papers, claims, scoring, and embedding config are recovered; generation-time metadata (exact git commit at gen time, host, exact command) is not recoverable.

1Source papers7
2Extracted claims7

Each row is one structured triple extracted by deepseek-v4-flash (prompt v1_2026_05_07, source: configured). Confidence is the LLM's self-rating.

  • bevacizumabinhibitsVEGFA→B
    "glioblastoma mouse model"confidence 0.90· extraction run #7· claim #1415
  • bevacizumabinhibitsVEGFA→B
    "in glioblastoma"confidence 0.80· extraction run #9· claim #7551
  • Bevacizumabbinds_toVEGFA→B
    "in immunoniosomes for glioblastoma treatment"confidence 0.90· extraction run #9· claim #9630
  • Bevacizumabbinds_toVEGFA→B
    "general mechanism"confidence 0.90· extraction run #19· claim #85832
  • bevacizumabinhibitsVEGFA→B
    "in newly-diagnosed glioblastoma patients after radiotherapy"confidence 0.95· extraction run #19· claim #87419
  • VEGFassociated_withAlzheimer's diseaseB→C
    "plasma levels comparing MCI individuals and Alzheimer's disease patients"confidence 0.90· extraction run #16· claim #46475
  • VEGFassociated_withAlzheimer's diseaseB→C
    "cerebrospinal fluid"polarity: negativeconfidence 0.90· extraction run #16· claim #58335
3Bridging logicSwanson ABC
Drug
Bevacizumab
binds_to
5 papers
Mediator
VEGF
associated_with
2 papers
Outcome
Alzheimer's disease
edge_support_log_ab
log(ab + 1) = 1.792
edge_support_log_bc
log(bc + 1) = 1.099
novelty_factor
1.000 (no direct drug→outcome evidence — full novelty)
direct_a_to_c_count
0
4Embedding modelbge-m3

Embeddings are used at the L1 filtering stage to prune off-topic abstracts before claim extraction. They do not directly score the hypothesis.

model
bge-m3
dim
1024
max_length
8192 tokens
similarity_scores
No per-hypothesis cosine scores stored.
5Final scoringhigh · 0.289
method
swanson_abc_log_evidence_with_novelty_penalty
formula
score = min((log(ab+1) + log(bc+1)) * novelty / 10, 1.0); novelty = 1.0 if direct==0 else 0.7 / (direct+1)
raw score
0.2890
outreach score
0.2890
quality tier
high
Pipeline version
9996ffa
Pipeline run
id 4 · uuid eb8b7730…
Pipeline command
scripts/backfill_provenance_v2.py:v2_2026_05_12
Pipeline status
completed
Generated
2026-05-12 21:43:39 UTC
Schema version
1.0
Download full provenance JSON
Score trajectory

How this hypothesis evolved

First captured May 11, 2026, latest May 13, 2026. 3 snapshots.

0.000.100.200.300.4002468 score evidence count May 11, 2026 · score 0.289 · 7 evidence · CT: none · Lit: partial_evidenceMay 11freezeMay 12, 2026 · score 0.289 · 7 evidence · CT: none · Lit: partial_evidenceMay 12freezeMay 13, 2026 · score 0.289 · 7 evidence · CT: none · Lit: partial_evidenceMay 13monthly
Score (left axis) Evidence count (right axis, dashed)