Abstract
<title>Abstract</title> <p> Taxonomically restricted genes (TRGs) are lineage-specific innovations, but in non-model genomes, their discovery is inflated by draft-assembly fragmentation and homology-detection failure (HDF). Across nine assemblies of the <italic>Anopheles gambiae</italic> sensu lato complex, the main malaria vectors, we build an assembly-artifact and HDF-resistant catalog of TRGs and candidate <italic>de novo</italic> genes. Because raw counts scaled with both fragmentation and proteome size, we based all conclusions on cross-species complex-core families annotated in at least three species: 213 high-confidence families, including a domainless subset of 116 candidate <italic>de novo</italic> gene families. These loci are transcribed, constrained, and not derived from transposons or annotations. Intrinsic disorder arises from conserved genes through <italic>de novo</italic> candidates to intergenic ORFs: as the pre-adaptation model predicts; among confidently folded models, no <italic>de novo</italic> candidate matched a known structure, whereas all housekeeping controls did, with characterized TRGs intermediate. The result is a validated catalog built on a portable, control-anchored framework. </p>