Reviewed twelve direct-source annotations across three complete reference groups. The source guide records original Methods/Results and figure-caption locators; the decision table retains full source joins.
The evidence supports focal-adhesion localization and β1-integrin activation in human keratinocytes. Two generic cell-adhesion terms are refined to the measured cell-matrix attachment role. Proliferation, migration, polarity and gap closure remain positive non-core cellular outcomes. Kindlin-2 compensation and its greater contribution to cell–cell contacts are preserved. Cytosol includes complexes; basal junction localization does not prove a hemidesmosomal or cell–cell-junction component.
The immutable 2011 cache omits operative Results despite its full-text flag; the original PMC body supplied those sections. The 2006 cache is abstract-only; author-uploaded original text and indexed publisher passages supplied the bounded evidence. Actual failed HTTP responses remain private provenance. Public descriptions of access do not imply a local PDF or independent pixel inspection.
The accepted intake records genuine provider failures and no generated research report. These notes are authored curation notes, not a provider report. All 31 imported source objects, nine reference identities/findings, four splice-product objects and the nineteen unowned reviews remain unchanged except for the three explicitly supplied whole-reference manual assessments. No new quotation, annotation, product, core function or top-level description is added in this lane. The 74/78-kDa experimental forms are not mapped to natural splice products.