Quickly verify claims by accessing the underlying experimental data and figures.
Press Enter ↵ to solve
Fuel Your Discoveries
"I would rather have questions that can't be answered than answers that can't be questioned."
- Richard Feynman
Quick Explanation
Copied
Quick critique — evidence-first summary
Mitra et al. compiled SBAs for 47 US‑approved peptide drugs (46 analyzed) and extracted reproducible patterns: most peptides followed small‑molecule toxicology paradigms (rats + dog/monkey), NPAA use was common (21/47), immunogenicity was frequent clinically (≈87% of tested drugs), genotoxicity testing was broadly negative, and carcinogenicity signals in rodents often reflected pharmacology rather than genotoxicity — highlighting inconsistent regulatory expectations and the need for peptide‑specific guidance
Long Explanation
Visual paper analysis — Development of peptide therapeutics: A nonclinical safety assessment perspective
Data source: FDA SBAs for US‑approved peptides (1998–2019) compiled by the authors (47 peptides identified; 46 included in most analyses)
Source and counts from SBAs compiled by authors; angiotensin II excluded from main analyses due to sparse toxicology data
Rats were used in 89% of programs (≈41/46); dogs and cynomolgus monkeys split non‑rodent testing
Authors report ADA assessed in 31/46 peptides (67%) overall; among those assessed, nonclinical ADA positive rate ≈62% and clinical ADA positive rate ≈87%; nonclinical ADA predicted clinical negativity with high specificity but had limited sensitivity for clinical positives
39/46 peptides had some genotoxicity testing; two Ames positives (glucagon and etelcalcetide) were resolved by follow-up analyses (glucagon false positive due to liberated histidine/tryptophan artifact; etelcalcetide cleared by mammalian assays), and overall peptides were considered non‑genotoxic in the dataset
About half the peptides had carcinogenicity studies (23/46) and ~57% of those were positive (13/23), often driven by pharmacology (e.g., GLP‑1 class thyroid C‑cell tumors in rodents), not genotoxic mechanisms
Key evidence-based takeaways
Regulatory heterogeneity: most sponsors used ICH M3(R2)‑style toxicology programs even for recombinant peptides, producing long chronic study durations (rodent ≥6 months; non‑rodent ≥9 months) rather than ICH S6(R1)-style biologic paradigms
Immunogenicity: clinical ADA is common (high percentage of positivity), but in most cases ADAs lacked measurable PK/efficacy impact in the available SBA data — however testing sensitivity and neutralizing assessments were often missing
NPAA and impurities: NPAAs were present in ~45% of peptides; sponsors rarely tested standalone NPAA toxicology unless NPAAs appeared as impurities — impurity thresholds and redacted levels in SBAs leave gaps in transparency
Genotoxicity: majority tested and negative; positive Ames results were explained by assay artifacts or resolved by mammalian tests — supports targeted rather than blanket genotox testing for peptides lacking cell permeability
Critical blindspots, biases & limitations in the paper (and dataset)
Dependence on SBAs: SBA documents are variably redacted and inconsistent in detail (e.g., impurity thresholds redacted), limiting transparency and reproducibility of extracted metrics
Temporal bias: ADA assay sensitivity and genotoxicity methodologies have evolved over decades; older approvals in dataset may under-report ADA or use less sensitive ADA assays, biasing longitudinal inferences
Species pharmacology gaps: many SBAs omitted explicit rationale tying toxicology species choice to pharmacological relevance — raises risk of rodent pharmacology-driven findings (e.g., rodent tumors) that are not human‑relevant
Conflict of perspective: all authors are industry-employed (Genentech, Ra Pharma, PTC) — paper is dataset-driven and transparent, but industry employment may influence emphasis and recommendations; authors disclose employment and claim no competing interests
Practical recommendations (evidence-weighted)
Adopt peptide‑specific regulatory expectations that: align species selection with pharmacology, avoid unnecessary genotoxicity batteries for peptides without cell permeability/NPAA risk, and clarify impurity/NPAA thresholds and qualification strategies
Require sensitive, validated ADA assays and neutralizing assessments when ADA could impact PK/PD or safety (e.g., analogs of endogenous proteins), and standardize reporting in SBAs to improve transparency and cross-program comparisons
Use mechanism‑driven carcinogenicity decisions: test rodents only when pharmacology is present in the species or when lifetime exposure and mechanism justify it; otherwise justify omission and propose alternative approaches (pharmacodynamic biomarkers, mode‑of‑action studies)
(Launch iterative agent to deep-dive SBAs, extract redacted thresholds, or programmatically re-analyze ADA/genotox metadata.)
Feedback:
Updated: February 17, 2026
BGPT Paper Review
Study Novelty
70%
The paper compiles and re-analyses a focused regulatory dataset (SBA packages) across three decades to derive practical safety‑assessment patterns for peptides; novelty arises from the dataset synthesis and regulatory framing rather than discovery of new biological mechanisms.
Scientific Quality
80%
High quality dataset work: transparent inclusion criteria (FDA SBAs 1998–2019), quantitative tallies (n=47), and careful endpoint summaries; limitations: dependence on variably redacted SBAs, potential author industry perspective bias (authors employed by industry), and descriptive (not experimental) design.
Study Generality
70%
Findings generalize across many peptide modalities and manufacturing methods (synthetic, recombinant, semi-synthetic) and have regulatory relevance, but are limited to US approvals and SBAs available by 2019.
Study Usefulness
80%
Practical and actionable: clarifies testing patterns and gaps (immunogenicity, NPAA, genotoxicity, carcinogenicity) that can inform sponsors and regulators and supports calls for peptide‑specific guidance.
Study Reproducibility
60%
Methods are described (data extraction from SBAs) but reproducibility is hampered by redacted SBA content and the evolving online availability of older SBAs; a programmatic re-extraction would assist reproducibility.
Explanatory Depth
80%
The authors give mechanistic/interpretive depth for key observations (e.g., Ames false-positive mechanism with liberated amino acids; pharmacology-driven rodent carcinogenicity) and discuss regulatory implications, though no new mechanistic experiments were performed.
Programmatically parse FDA SBA PDFs, extract structured fields (manufacturing method, species used, ADA results, genotox results, NPAA presence), and output a reproducible dataset for meta-analysis and visualization (SBA to CSV).
Get emailed when your analysis is done!
We'll email you the results when your analysis is finished.
Hypothesis Graveyard
Routine Ames testing for peptides containing proteogenic amino acids — rejected because false positives (e.g., liberated histidine/tryptophan) and lack of cell penetration make bacterial assays poor predictors.
Blanket two‑rodent carcinogenicity studies for all peptides — rejected because many rodent tumors reflect exaggerated pharmacology in species not relevant to human exposure and thus do not predict human carcinogenic risk.
Quick Explanation
Long Explanation
Top Data Sources
ExportMCP
Keep Exploring
Analysis Wizard
Hypothesis Graveyard
Potential Experiments
Discussion
Fuel Your Discoveries
Ask a Follow-Up
Key Insight
