COMPOUND DEEP DIVES · PEPTIDE SCIENCE 101
Most researchers assume a Certificate of Analysis is a rubber stamp — a document that confirms what the label says. That assumption is wrong, and the analytical literature is unambiguous about why.
A Certificate of Analysis (CoA) is only as reliable as the method used to generate it. Standard single-dimension reversed-phase HPLC — the platform cited on the overwhelming majority of commercially available research peptide CoAs — is consistently shown in peer-reviewed chromatographic literature to fail at resolving structurally similar impurities: d-/l-enantiomers, single-amino-acid deletion sequences, truncated chains, and closely eluting substitution variants all share enough structural similarity with the target compound that they co-elute under generic C18/TFA gradient conditions and inflate the reported purity figure (Karongo et al., 2020, PMID: 32823119; Stoll et al., 2023, PMID: 36871316). The gap between “reported purity” and “true purity” is, in most commercial CoA contexts, an unknown quantity.
This matters most for researchers working with bioactive peptides — BPC-157, TB-500, Epithalon, CJC-1295, GHK-Cu — where small structural deviations from the target sequence can meaningfully alter receptor binding profiles, change the pharmacodynamic signal, or introduce uncharacterised impurities that confound experimental results. The analytical standards applied to pharmaceutical peptides under ICH guidance are not automatically applied to research-grade suppliers. The difference shows up on the CoA — if you know what to look for.
This post decodes every section of a research-grade peptide CoA, explains the analytical method behind each number, names the specific failure modes in commonly used approaches, and tells you exactly what red flags to look for before accepting a purity claim.
The analytical chemistry of peptide quality control is a mature field, with validated methods established for pharmaceutical compounds including atosiban, teduglutide, leuprolide, oxytocin, and exenatide. The research reviewed here draws on in vitro characterisation studies conducted under controlled laboratory conditions, using reference-grade instruments (UHPLC-Orbitrap MS, ESI-QTOF, UPLC-ToF, SFC-MS) that represent the upper end of what the field can achieve.
The core challenge in peptide purity assessment is the structural similarity of synthetic impurities to the target molecule. Solid-phase peptide synthesis (SPPS) — the manufacturing method for virtually all research-grade synthetic peptides — generates characteristic impurities at each coupling step. Deletion sequences arise when one coupling reaction fails and synthesis continues; the result is a peptide one amino acid shorter than the target. Truncated sequences occur when chain elongation terminates prematurely. Amino acid substitution variants arise from racemisation or cross-contamination during synthesis. These impurities are not random — they are structurally predictable — which is why validated stability-indicating HPLC methods can and should separate them. The question is whether the CoA you receive was generated using such a method.
Karongo et al. (2020, PMID: 32823119) developed a comprehensive reversed-phase × reversed-phase 2D-LC system (acidic first dimension, basic second dimension, with complementary UV at 215 nm, charge-aerosol detection, and ESI-QTOF-MS) and applied it to exenatide, octreotide, cyclosporine A, and oxytocin. The critical finding: the 2D system resolved structurally similar impurities — including d-/l-isomers — that co-eluted and were therefore invisible in conventional 1D-RPLC runs of the same compounds.
Stoll et al. (2023, PMID: 36871316) formalised this with a two-step scouting procedure to derive optimal second-dimension gradient conditions for peak purity assessment in pharmaceutical peptides. The authors note directly that QC methods for pharmaceutical peptides “almost universally depend on reversed-phase liquid chromatography” and that co-eluting impurities represent a “significant but underappreciated risk.” This is a statement from the pharmaceutical QC literature about the best-regulated segment of the market; for research-grade suppliers, the risk is higher.
Pérez-Robles et al. (2022, PMID: 36263764) validated a combined UHPLC-UV and Orbitrap MS method for teduglutide, demonstrating that forced degradation (thermal stress at 40°C and 60°C, mechanical shaking, photolytic exposure) generated products detectable only by MS — not UV alone. Pawar et al. (2024, PMID: 38689387) applied an ICH-validated gradient RP-HPLC method to a novel synthetic decapeptide and characterised six distinct synthesis- and storage-stage impurities under identical stress conditions, establishing that each impurity profile must be individually characterised by name or structure — not listed generically as “related substances.”
The purity percentage on a peptide CoA is almost always derived from RP-HPLC area-percent analysis: the instrument integrates the UV absorbance at 215–220 nm across the entire chromatogram, and purity is calculated as (area of main peak ÷ total integrated area) × 100. This method works because the peptide bond absorbs UV light at that wavelength, making it effectively universal for peptide-containing species.
The problem is selectivity. When the main peak and an impurity have similar enough structures — a deletion sequence that differs by one residue, an enantiomeric d-amino acid substitution — they can co-elute under standard C18/TFA gradient conditions. The integrator sees one peak. The purity figure reads 98%. The actual content of the target peptide may be 91%.
Schleiff et al. (2024, PMID: 39068812) demonstrated this precisely with glutathione: an FDA laboratory developed an HPLC-UV method achieving LOD 0.02% w/w and LOQ 0.05% w/w for four related impurities and a newly identified degradant (L-pyroglutamic acid). Applied to commercial bulk glutathione samples, several showed quantifiable impurities not detectable by the conventional iodine titration method used for historical CoA generation. Different method, different purity figure — on the same physical material.
| Compound | Study Type | Key Outcome | Citation |
|---|---|---|---|
| Glutathione (tripeptide) | In vitro HPLC-UV validation | Commercial bulk samples showed quantifiable impurities (LOD 0.02% w/w) not detectable by conventional iodine titration; selective method revealed impurity profiles obscured by non-selective assay | Schleiff et al., 2024, PMID: 39068812 |
| Oxytocin | In vitro 2D-LC (acidic × basic RP) | d-/l-isomers and amino acid substitution variants resolved in 2D but co-eluted in 1D-RPLC; standard CoA purity figures for oxytocin unreliable for structural impurities | Karongo et al., 2020, PMID: 32823119 |
| Exenatide / Exendin-4 | In vitro 2D-LC + ESI-QTOF-MS | Co-eluting impurities invisible to 1D-RPLC resolved by second-dimension orthogonal separation; UV area-percent overestimates target compound purity | Karongo et al., 2020, PMID: 32823119 |
| Teduglutide (GLP-2 analogue) | In vitro UHPLC-UV-Orbitrap MS, forced degradation | Degradation products from thermal and photolytic stress detectable only by MS, not UV alone; UV-only CoA cannot confirm absence of degradants | Pérez-Robles et al., 2022, PMID: 36263764 |
| Novel synthetic decapeptide | In vitro gradient RP-HPLC, ICH validation | Six distinct synthesis- and storage-stage impurities characterised; diverse impurity profiles require named identification, not generic “related substances” listing | Pawar et al., 2024, PMID: 38689387 |
Red flag: A CoA that reports a single purity figure (e.g., “≥98% by HPLC”) without specifying the column type, mobile phase modifier (TFA vs. formic acid), gradient programme, and detector wavelength is not providing enough information to assess whether the method was capable of resolving the specific impurities that SPPS generates for that compound class.
MS identity confirmation is the second major section of a quality CoA, and it is where most commercial documents fall short. “Confirmed by mass spectrometry” as a standalone phrase is analytically indefensible. Here is why.
ESI-MS ionises peptides in positive-ion mode, generating charged adducts. Smaller peptides (<1,500 Da) typically generate singly charged [M+H]⁺ ions. Larger peptides generate multiply charged ions: [M+2H]²⁺, [M+3H]³⁺, and so on. Zajickova et al. (2020, PMID: 32867492) confirmed the identity of gramicidin pentadecapeptides (A, B, and C) via ESI-MS by detecting specific adduct combinations ([M+2H]²⁺, [M+2Na]²⁺, [M+H+Na]²⁺) in positive-ion mode within a 3-minute SFC-MS run. The key point: each detected m/z value maps back to a specific theoretical molecular weight within ≤0.1 Da on high-resolution instruments.
Skiba et al. (2020, PMID: 32920482) validated an LC-MS/MS method for the GnRH analogue leuprolide using specific MRM transitions (m/z 605.5→110.2), demonstrating that MS identity confirmation at the highest confidence level requires reporting the precursor-to-product ion transition — not just the precursor mass.
A rigorous CoA MS section should report:
– Theoretical monoisotopic molecular weight of the target peptide (e.g., 1,419.58 Da for BPC-157)
– Observed m/z value(s) with charge state notation (e.g., [M+2H]²⁺ observed at m/z 710.84)
– Instrument type (quadrupole, Orbitrap, ToF — resolution matters)
– Agreement within ≤0.1 Da (high-resolution) or ≤1 Da (low-resolution) of theoretical
| Compound | Study Type | Key Outcome | Citation |
|---|---|---|---|
| Leuprolide (GnRH analogue) | In vitro LC-MS/MS validation | Identity confirmed via MRM transition m/z 605.5→110.2; validates that MS identity on CoA requires specific ion transitions, not just MW match | Skiba et al., 2020, PMID: 32920482 |
| Gramicidin A/B/C (pentadecapeptides) | In vitro SFC-ESI-MS | Full separation of isoforms in 3 min; identity via [M+2H]²⁺, [M+2Na]²⁺, [M+H+Na]²⁺; demonstrates expected multiply-charged ion patterns for peptides >1,000 Da | Zajickova et al., 2020, PMID: 32867492 |
| Atosiban (nonapeptide) | In vitro RP-HPLC + LC-MS | Five synthesis impurities separated and characterised by MW; demonstrates that MS on a CoA should identify impurity MW, not just confirm main compound | Li et al., 2020, PMID: 31585328 |
| Fc fusion protein truncations | In vitro UPLC-ToF MAM, 3-lab co-validation | Multi-attribute MS method confirms identity AND quantifies truncation variants per ICH Q2(R1); establishes that MS CoA data should include CQA quantification, not identity only | Wu et al., 2023, PMID: 36372921 |
Red flag: A CoA that states “Identity: Confirmed” or “MS: Pass” without an observed m/z value is providing qualitative confirmation at best. Without the reported mass value, you cannot verify independently that the spectrum was run on the correct compound.
TFA (trifluoroacetic acid) is the standard ion-pairing agent in RP-HPLC peptide purification. It binds tightly to basic amino acid residues (lysine, arginine, histidine) as a counterion salt and remains associated with the peptide after lyophilisation unless a specific desalting or counter-ion exchange step is performed.
Streuli et al. (2026, PMID: 41667417) demonstrated that the choice of mobile-phase modifier has direct consequences for purification quality and residual content: substitution of TFA with formic acid improved separation performance in specific cases, and they developed a correction equation that reduced method-transfer errors from ~17% to <5%. Govender et al. (2020, PMID: 32871421) showed that SFC purification using CO₂/methanol systems leaves no TFA residue — an advantage over conventional RP-HPLC in terms of counterion profile.
The analytical problem: TFA has a molecular weight of 114 Da. On a peptide like Semax (MW ~800 Da), significant TFA counterion loading can inflate the apparent peptide mass by 10–30% if not removed and corrected. A CoA that reports purity as area-percent by HPLC but does not separately report TFA or acetate counterion content — ideally by ion chromatography or ¹⁹F NMR — cannot confirm what percentage of the reported mass is actually the target peptide.
What to look for: A counterion/salt content specification, ideally as TFA ≤X% (w/w) or “acetate form” with corresponding analytical method noted.
HPLC purity and MS identity tell you about chemical composition. They tell you nothing about microbial contamination. Bacterial endotoxins (lipopolysaccharides) are biologically active at concentrations in the picogram-per-millilitre range in many in vitro and animal model contexts, and they are completely invisible to UV-based HPLC.
The standard test is the LAL (Limulus Amebocyte Lysate) assay, which detects endotoxin at ≤0.1 EU/mg in most pharmaceutical standards. Absence of an endotoxin section on a research-grade CoA is a meaningful red flag, particularly for compounds intended for cell-based or animal model research where LPS contamination is a known confound.
Lambert et al. (2018, PMID: 30202518) surveyed real-world oxytocin ampoules from healthcare facilities across five provinces in the Democratic Republic of Congo. 80% of 15-facility samples contained less than 90% of specified peptide content when re-tested by validated HPLC and LC-MS; all samples contained an unidentified impurity averaging 12.3% of the main peak (range: 8.0–20.5%). The study is the most direct real-world evidence available that nominal label claims and non-validated vendor documents do not reliably represent true analytical purity.
Every production batch from an SPPS manufacturer has a unique impurity profile. Raw material lot, operator, coupling reagent freshness, and resin loading efficiency all vary between runs and generate batch-to-batch differences in impurity fingerprint. A CoA that lacks a lot number cross-referenced to the received material — or uses the phrase “representative certificate” — is describing a different batch of material than the one you are holding.
ICH Q2(R1) defines the validation requirements that make an analytical method defensible: specificity, linearity (R² >0.999), LOD, LOQ, precision (CV% <2%), accuracy, and robustness. Wu et al. (2023, PMID: 36372921) demonstrated in a three-laboratory co-validation study that even rigorous MAM methods required careful design to achieve all pre-defined analytical target profiles — and still found deviations >4% in specific sample matrices. A single-laboratory CoA from a commercial research supplier that does not disclose validation status has an inherently uncharacterised precision range.
The complete CoA checklist for research-grade peptides:
| CoA Section | Minimum Requirement | Red Flag |
|---|---|---|
| HPLC Purity | Method specified (column, mobile phase, gradient, wavelength); ≥95% area-percent; stability-indicating validation noted | “≥98% by HPLC” with no method details |
| MS Identity | Observed m/z with charge state; theoretical MW stated; instrument type noted; agreement ≤0.1 Da (HR) | “Confirmed by MS” with no m/z value |
| Related Substances | Individual impurities named or structurally characterised; limits stated per impurity | Generic “related substances: NMT 2%” |
| TFA/Counterion | TFA content stated (% w/w) or explicit note that acetate exchange was performed | No mention of counterion content |
| Endotoxin | LAL result in EU/mg; test method stated | Section absent entirely |
| Water Content | Karl Fischer result (% w/w) | Section absent |
| Lot Number | Batch-specific number; date of analysis | “Representative CoA” or no lot number |
| Method Validation | ICH Q2(R1) reference or equivalent; validation parameters listed | No validation status disclosure |
The analytical literature reviewed here is clear on the technical inadequacy of standard single-dimension RP-HPLC for comprehensive peptide purity assessment. That conclusion deserves some qualification before applying it to research-grade sourcing decisions.
Limitation 1: Pharmaceutical ≠ Research-Grade Standards
All studies reviewed were conducted on pharmaceutical-grade compounds — atosiban, leuprolide, teduglutide, oxytocin, exenatide — using reference-standard instruments under ICH-guideline validation frameworks. The analytical precision achievable in these settings (Orbitrap MS at ≤0.001 Da mass accuracy, 2D-LC systems with CAD and QTOF detection) significantly exceeds the equipment typically deployed by commercial research-grade peptide suppliers. The literature tells us what ideal CoA data should look like; it does not tell us how closely any given commercial supplier approaches that ideal, because commercial CoAs are not subject to independent audit. The gap between what the science demands and what is practically disclosed varies by vendor.
Limitation 2: Co-Elution Risk Is Real But Compound-Specific
Karongo et al. (2020) and Stoll et al. (2023) both document co-elution failures in specific peptide systems under specific gradient conditions. The severity of co-elution risk depends on the structural similarity between the target compound and its SPPS-derived impurities — a glycine-deleted sequence from a peptide with terminal glycine will behave differently than an internal deletion. The published literature cannot be directly extrapolated to every research-grade compound without compound-specific orthogonal testing. What can be stated is that the risk is non-zero and is not eliminated by reporting 1D-RPLC purity alone.
Limitation 3: Purity Thresholds Are Convention, Not Biology
The commonly cited research-grade purity thresholds — ≥95% or ≥98% — derive from pharmaceutical industry convention, not from controlled dose-response studies linking a specific purity level to differential biological outcome in a given model. The literature does not currently provide data on what purity delta produces a detectable change in in vitro receptor binding or animal model response for any of the specific compounds in the research compound catalogue. Whether the difference between 95% and 98% purity is functionally meaningful depends on the identity of the impurities at the remaining 5% or 2% — which requires characterised related substances data to assess.
Limitation 4: A CoA Is a Point-in-Time Document
Forced degradation studies (Pérez-Robles et al., 2022; Pawar et al., 2024) demonstrate that synthetic peptides degrade under thermal, oxidative, and photolytic stress, generating distinct impurity profiles. A CoA reflects the analytical state of the material at the time of testing. Without real-time stability data under the specific storage conditions applied after receipt — and without a re-test commitment at defined intervals — the CoA cannot guarantee the purity of the material at the point of use. This is an underappreciated gap in most commercial CoA frameworks.
Limitation 5: Interlaboratory Variability Is Uncharacterised for Single-Lab CoAs
Wu et al. (2023, PMID: 36372921) demonstrated that three independent laboratories following the same validated protocol still produced quantification deviations in specific sample matrices. A single-laboratory CoA — which describes all commercial research-grade CoAs — has an inherently uncharacterised inter-laboratory precision component. This does not invalidate single-lab data; it means the precision range of that data is unknown unless the vendor discloses instrument calibration records, analyst qualification documentation, and round-robin proficiency data.
Limitation 6: Absence of Human Safety Data on Impurity Thresholds
No human pharmacokinetic or safety studies specifically link CoA purity thresholds to differential biological outcomes. This is a fundamental limitation of the current literature and a reason why the preclinical framing of all compound research — including the analytical chemistry supporting it — remains the appropriate epistemic frame.
Reading a peptide CoA is a practical analytical literacy skill. The research is unambiguous that the purity figure — taken alone, without method context — is insufficient to characterise what you are actually working with.
The minimum defensible CoA for a research-grade peptide includes: a specified and validated HPLC method with column and mobile phase disclosed; an MS identity confirmation that states observed m/z and charge state, not just “confirmed”; an individual related-substances breakdown with characterised impurity identities; a TFA or counterion content figure; an LAL endotoxin result; Karl Fischer water content; and a batch-specific lot number cross-referenced to the analytical date.
Compounds across the research notes and product range at biohacker.team — from recovery compounds like TB-500 and BPC-157 to cognitive compounds like Semax and Pinealon, longevity compounds like Epithalon and GHK-Cu, and metabolic compounds like CJC-1295 and Tesamorelin — all exist in a sourcing environment where analytical quality varies significantly. The CoA is the primary instrument for assessing that quality. This post gives you the vocabulary to use it correctly.
The practical takeaway for researchers: when evaluating any peptide source, request the full CoA before purchase and run it against the checklist in Section 4 above. If a supplier cannot or will not provide method details alongside a purity figure, the purity figure is not analytically defended. That is not a regulatory opinion — it is what the chromatographic literature says.
Karongo R et al. (2020). A selective comprehensive reversed-phase×reversed-phase 2D-liquid chromatography approach with multiple complementary detectors as advanced generic method for the quality control of synthetic and therapeutic peptides. Journal of Chromatography A. PMID: 32823119.
Stoll DR et al. (2023). A Strategy for assessing peak purity of pharmaceutical peptides in reversed-phase chromatography methods using two-dimensional liquid chromatography coupled to mass spectrometry. Part II: Development of second-dimension gradient conditions. Journal of Chromatography A. PMID: 36871316.
Streuli A et al. (2026). Improvement of Analysis and Transferability in Peptide Purification: From HPLC to FPLC and Back Again. Journal of Peptide Science. PMID: 41667417.
Pérez-Robles R et al. (2022). Method for identification and quantification of intact teduglutide peptide using (RP)UHPLC-UV-(HESI/ORBITRAP)MS. Analytical Methods. PMID: 36263764.
Pawar R et al. (2024). A stability-indicating method development and validation for the determination of related substances in novel synthetic decapeptide by HPLC. Journal of Peptide Science. PMID: 38689387.
Schleiff M et al. (2024). An isocratic HPLC-UV analytical procedure for assessment of glutathione and its related substances. Journal of Pharmaceutical and Biomedical Analysis. PMID: 39068812.
Li Q et al. (2020). A novel HPLC method for analysis of atosiban and its five related substances in atosiban acetate injection. Journal of Pharmaceutical and Biomedical Analysis. PMID: 31585328.
Lambert P et al. (2018). Quality of oxytocin ampoules available in health care facilities in the Democratic Republic of Congo: an exploratory study in five provinces. Journal of Global Health. PMID: 30202518.
Wu G et al. (2023). Interlaboratory Co-validation of a UPLC-ToF MS MAM Method for Truncations of a Fc Fusion Protein. Current Pharmaceutical Biotechnology. PMID: 36372921.
Skiba M et al. (2020). Development, validation and method stability study of a LC-MS method to quantify leuprolide (Hormone analog) in human plasma. Journal of Chromatography B. PMID: 32920482.
Zajickova Z et al. (2020). Monolithic Poly(styrene-co-divinylbenzene) Columns for Supercritical Fluid Chromatography-Mass Spectrometry Analysis of Polypeptide. Analytical Chemistry. PMID: 32867492.
Govender K et al. (2020). The development of a sub/supercritical fluid chromatography based purification method for peptides. Journal of Pharmaceutical and Biomedical Analysis. PMID: 32871421.
Every compound supplied by biohacker.team is accompanied by a batch-specific Certificate of Analysis generated from HPLC purity data at ≥215 nm with method parameters disclosed, ESI-MS identity confirmation with observed m/z and theoretical molecular weight stated, and a named related-substances section. COA documents are available for download on each product page; HPLC chromatograms and MS spectra are available on request. Our sourcing partners operate under cGMP-aligned quality frameworks with HPLC instrument calibration logs maintained per ICH Q2(R1) requirements. We do not issue representative or generic CoAs — every document is batch-specific and cross-referenced to the lot number of the material supplied. You can review our full sourcing transparency position at biohacker.team/about/.
For research use only. Not for human consumption. Not intended to diagnose, treat, cure, or prevent any disease or condition. All compounds are sold strictly for in-vitro and animal research purposes. Not approved for human use.