A new study combines existing genetic scores using local biobank data. Our guide separates model ranking, probability calibration and clinical usefulness, while recording a numerical discrepancy between the accepted HTML and PDF summaries.
Methods explainer: prediction models, validation and source versions
A genetic reconstruction finds earlier epidemic activity and a changing association with acute flaccid myelitis. Our guide separates sequence sampling, inferred transmission and reported neurological cases.
6 min read · 23 Oct 2026
Understanding evidence
Methods explainer: metagenomic catalogues and ecological inference
A global livestock-manure study separates resistance genes annotated in viral sequences from predicted links between phages and resistant bacteria. Those are different kinds of evidence—and different denominators.
6 min read · 23 Oct 2026
Understanding evidence
Methods explainer: receptor discovery and evidence across biological scales
A new orthobunyavirus study identifies a cellular entry factor and compares related viruses. We explain why a cell-entry result, an animal outcome and real-world transmission answer different questions.
6 min read · 23 Oct 2026
Understanding evidence
Methods explainer: biochemical reconstitution and single-molecule measurements
Purified yeast proteins reveal how a temporary DNA bridge can delay end processing. The key control separates physical assembly from successful joining.
6 min read · 23 Oct 2026
Understanding evidence
Methods explainer: bacterial regulatory networks and targeted validation
A map of 370 transcription factors links DNA binding, growth-stage expression and strain diversity. Our evidence guide separates a predicted connection from a tested biological effect.
Live-cell mapping places eleven FG nucleoporins into a layered architecture. We explain what the reconstruction measures, how another method checks it and why a preferred location is not a fixed structure.
A new genetic screen finds compensating changes in pneumococcus. The useful result is not that essential genes are unimportant, but that their necessity depends on biological context.
A reanalysis of 4,839 public stool metagenomes asks which associations persist across time and bacterial strains. Our guide keeps specimens, people and inflammatory state separate.
6 min read · 22 Oct 2026
GO BEYOND THE HEADLINE
The claim. The measurement. The missing link.
Follow the numbers to their source, compare the actual experiments and download the evidence notes.
