GSE147863

This is a bulk human RNA-seq dataset that earns a clean A: every metric feeding the score was directly measured (evidence_strength=1), so the grade rests on real QC output rather than extrapolation. The two strongest drivers are the base-call quality and library complexity — 98.9% of bases at Q30 (mean quality 35) means base errors are vanishingly rare and won't seed spurious variants or quantification noise, while a 1.54% duplication rate plus 0% adapter contamination indicate a complex, well-prepped library with little PCR or trimming artifact. For reuse the main caveat is not a quality failure but read geometry: a 28 bp mean read length is short for RNA-seq and will limit multi-mapping resolution, splice-junction detection, and isoform-level work, so this is most trustworthy for gene-level expression rather than transcript- or fusion-level analyses. Note also a 429 fetch error blocked some upstream metadata retrieval, but it did not affect the measured QC metrics, so the rating itself is firmly grounded.