Reproducibility: How Material Variability Undermines Experiments
Reproducibility has been examined intensively across the life sciences over the past decade1, and a recurring finding is that a substantial share of irreproducible results traces to reagent and material variability2 rather than to statistics or study design.
The variables that travel with a peptide
Two vials labelled with the same sequence can differ in purity, in the identity and profile of impurities, in salt form, in net peptide content, and in moisture. Each of these changes the actual quantity of active molecule delivered at a nominally identical concentration, and some change the biological background as well.
How the error propagates
Suppose a laboratory prepares a stock solution by weighing labelled milligrams and assuming the solid is entirely peptide. If net peptide content is materially below 100%, every downstream concentration is overstated by the same factor. Compare that to another laboratory using a lot with different net content, and two groups studying the same compound at the same nominal concentration are studying different exposures. Neither is aware of the discrepancy, because both followed the label.
Impurities as confounders
Impurity profiles matter independently. Endotoxin can drive immune readouts on its own. Residual trifluoroacetate can affect viability in sensitive cell systems. A deletion-sequence impurity may retain partial activity at the same target, subtly shifting a dose-response curve without producing an obvious anomaly.
What good practice looks like
The mitigations are unglamorous and effective: record the lot number in the experimental record alongside the compound name, retain the certificate of analysis with the data, calculate concentrations from net peptide content rather than gross mass, and where a study spans a long period, note when a lot change occurred. Material provenance is experimental metadata, and treating it that way is what makes results checkable later.
References
- Baker M. 1,500 scientists lift the lid on reproducibility. Nature. 2016;533(7604):452–454. doi:10.1038/533452a
- Freedman LP, Cockburn IM, Simcoe TS. The economics of reproducibility in preclinical research. PLoS Biol. 2015;13(6):e1002165. doi:10.1371/journal.pbio.1002165
Disclosure. This article is educational and summarises published scientific literature. It is not medical advice, and it does not describe outcomes you should expect. These statements have not been evaluated by the Food and Drug Administration, and the products discussed are not intended to diagnose, treat, cure, or prevent any disease. The products discussed are supplied for laboratory and research use only and are not for human or veterinary use, administration, or consumption. For adults 21 and over.
