How to Report Negative Results in Regulatory Element Studies

Report four things and the negative result becomes defensible: the claim you tested, the smallest effect your design could detect, the controls that show the assay works, and the alternative explanations you excluded. Leave any of them out and "no change" reads as "we did not look carefully".

Reviewers do not reject negative findings. They reject under-powered ones, and the objection usually has one of four forms:

• • The assay may not have been sensitive enough.
• • The perturbation may not have worked.
• • The cell type may have been wrong.
• • A compensating element may have masked the effect.

Each objection has a standard answer, and the answers are cheap to produce if planned in advance. The purpose of this template is to make them available at submission.

The detectable effect row is the one that most often decides the outcome. It converts a null from an absence of evidence into a bounded statement.

The distinction is not pedantry. The first column makes a universal claim from a bounded experiment; the second makes a bounded claim, which is what the data support.

A negative result without a positive control in the same experiment is uninterpretable, because "no effect" and "broken assay" produce identical data. Three positive controls are worth carrying routinely:

• • A perturbation with a known large effect in the same cells (an essential enhancer, a core promoter motif).
• • A technical positive control for the readout itself (a spike-in, or a gene known to respond to the treatment).
• • A guide or edit efficiency control showing that the perturbation machinery functioned at the targeted site.

If the positive control works and the test element shows nothing, you have a result.

A well-reasoned negative result names its own limits. The most important limits in regulatory studies:

• • Redundancy. Single-perturbation nulls do not exclude compensation; state whether combination perturbations were tested.
• • Cell type. A null in one background says little about another; state the context and, if possible, test a second.
• • Timing. Developmentally timed elements may show effects only during a window.
• • Sensitivity floor. Small true effects below the detection limit are not excluded.

Naming these limits does not weaken the paper. It demonstrates that the null was interpreted rather than merely observed.

• • In the main text when the null is a substantive finding — for example, that a well-annotated element is dispensable in a disease-relevant cell type.
• • In supplementary data when it supports a specificity argument.
• • In a dedicated resource when the value is in the systematic sweep rather than the individual element.

Depositing the underlying data — guide counts, clone-level expression, raw reads — matters more for negative results than for positive ones, because the reader cannot otherwise audit the detection limit.

EDITGENE builds the perturbation panels that make negative results defensible: multiple independent clones per element, matched non-targeting controls, validated positive control elements, and measured assay variability reported alongside every line.

Tell us the element and the claim you expect to be null. Contact EDITGENE to request a project consultation or quote.

Reporting standards differ by journal and field; the template above describes a defensible minimum for regulatory element studies, not a formal guideline.

A bounded null remains a statement about specific conditions, cell type, and assay sensitivity, and should be cited as such.

• • Fanelli D. Negative results are disappearing from most disciplines and countries. Scientometrics. 2012;90:891–904.
• • Osterwalder M, et al. Enhancer redundancy provides phenotypic robustness in mammalian development. Nature. 2018;554:239–243.
• • Fulco CP, et al. Systematic mapping of functional enhancer–promoter connections with CRISPR interference. Science. 2016;354:769–773.
• • Lakens D, et al. Justify your alpha. Nature Human Behaviour. 2018;2:168–171.