Skip to content

URAT1 mRNA Target-Site Selection — COMP-009 Invalidated

COMP-009 does not establish a usable URAT1 siRNA guide, target-site tractability, accessibility, specificity, cross-species reuse, or support for H03. No candidate sequence, funnel count, rank, score, shortlist, GREEN verdict, or P2-2 closure survives. The artifact is a non-runnable, hash-bound tombstone.

Why the result is invalid

The rerun used RefSeq NM_144585.4, correcting the original artificial back-translated CDS. Its decision model remained invalid:

  • The Reynolds implementation applied cited sense-strand positional preferences to the antisense strand, omitted terminal-stability and inverted-repeat criteria, and substituted a four-base homopolymer check. Reynolds et al.
  • The Ui-Tei gate did not require the cited terminal-composition, A/U-richness, and long-GC-stretch conditions simultaneously. Ui-Tei et al.
  • RNAplfold accessibility had no acceptance threshold. Five shortlisted windows produced a GREEN verdict regardless of their accessibility values.
  • The composite score used uncalibrated weights and allowed protein conservation to dominate the accessibility term. It was not the calibrated and independently tested RNAxs method. Tafer et al.
  • No transcriptome or 3′-UTR off-target clearance was performed; only one transcript was scanned; protein conservation cannot establish cross-species guide reuse; and one boundary-spanning window was mislabeled by a midpoint rule.
  • No candidate was tested for intracellular activity, URAT1 knockdown, target-cell uptake, urate transport, or renal safety.

What survives

The historical rerun examined NM_144585.4, but that fact has no predictive or decision use. The URAT1-siRNA hypothesis survives independently of COMP-009.

Selective proximal-tubule delivery is the upstream gate. COMP-048 is designed to ask whether an internalizing surface receptor co-localizes with SLC22A12-positive human proximal-tubule cells selectively enough to justify receptor-targeted delivery work.

A new guide-design COMP is deferred until a delivery route survives. It must use a validated current design method, cover relevant SLC22A12 transcripts and human variation, perform transcriptome-wide off-target analysis, keep accessibility separate from other evidence dimensions, and require empirical URAT1 knockdown before a guide advances.