Evaluating a claim about a gene-edited crop
A company reports that a gene-edited wheat line yields 12% more grain than the conventional variety in field trials. The edit disables a gene that slows starch breakdown in the grain. Trials were run at three company sites over one season, with four plots per variety per site. The report states that yield differences were "statistically significant" but gives no error bars, no soil data and no information on whether plots were randomised. A newspaper article says the edit "boosts photosynthesis and will feed millions". An independent scientist asks to see the raw data and is refused.
1. Evaluate the newspaper's claim that the edit boosts photosynthesis, using the information given.
Read the worked answer
The claim is not supported. The described edit disables a gene affecting starch breakdown in the grain, which is a post-harvest or grain-quality trait, not a photosynthetic pathway. Nothing in the report links the edit to carbon fixation or light capture.
The 12% yield increase, even if real, could arise from altered starch partitioning, greater grain filling or reduced losses, none of which is photosynthesis. The newspaper has conflated increased yield with increased photosynthesis, which is a specific mechanism claim requiring measurements such as photosynthetic rate or Rubisco activity.
Additionally, the evidence is weak: one season, three company sites, four plots per variety, no soil data, no randomisation details and no error bars. The claim that it will "feed millions" is a prediction far beyond the data. The appropriate response is that the edit may affect grain composition and possibly yield, but the photosynthesis claim is unsupported and the food-security claim is speculative.
2. Explain what the refusal to release raw data means for how the evidence should be judged, and outline the checks a reviewer would want.
Read the worked answer
Meaning of refusal. Science as a Human Endeavour emphasises that scientific claims are evaluated through peer scrutiny and reproducibility. Refusing raw data prevents independent verification, so the claim cannot be treated as established even if the company's summary is accurate. Publication in a press release is not peer review.
Checks a reviewer would want.
- Plot randomisation and blocking, to exclude soil or position effects.
- Error bars and the statistical test used, since "significant" without a stated test or variance is uninformative.
- Replication across seasons and independent sites, because one season cannot separate genotype from weather.
- Soil nutrient and water data, since yield differences may reflect management.
- Measurement of the edited gene's expression and of grain starch content, to confirm the proposed mechanism.
- Assessment of off-target edits and of any regulatory or biosafety requirements.
Without these, the 12% figure is a preliminary result. The reviewer should suspend judgement rather than accept or reject the claim outright, and should note the conflict of interest created by company-run trials.