When Does Framework Topology Matter Beyond Pore Metrics in MOF Adsorption Prediction?

MOF chemists often evaluate frameworks through two complementary languages: familiar pore metrics and framework topology. In adsorption screening, however, fast predictive models usually describe a structure mainly by density, surface area, pore apertures, cavity sizes and void fractions. This raises a practical question: after these pore metrics are known, does topology still add useful information? Here, we compare topology-only, geometry-only and topology+geometry descriptors for methane and carbon dioxide uptake across low-and high-pressure conditions. Compact geometry is the essential first baseline and almost saturates high-pressure methane prediction; adding topology changes the HGB reference model R2 only from 0.974 to 0.977. Topology is therefore not a universal shortcut and cannot replace pore metrics. Its value is selective and becomes clearer when pore access, confinement and network connectivity influence adsorption, increasing R2 from 0.673 to 0.734 for methane at 5.8 bar, from 0.678 to 0.732 for carbon dioxide at 0.15 bar and from 0.463 to 0.536 for carbon dioxide at 0.015 bar. Matched-geometry comparisons, residual diagnostics, net-specific maps and carbon dioxide shortlist analyses show that topology can change which frameworks are selected even when scalar pore descriptors look similar. The conclusion is a practical design rule for MOF screening: pore metrics should be the first lens, but topology becomes a useful complementary descriptor when adsorption depends on how pore space is connected, not only on how much pore space is present.