An integrated platform for genome-wide functional analysis in Aspergillus niger through Ac/Ds insertion profiling and metabolic modeling

Genome-scale metabolic models guide rational strain engineering, while high-throughput insertional mutagenesis enables systematic investigation of gene function and condition-dependent growth requirements. However, a platform integrating high-density transposon insertion sequencing with matched stoichiometric, thermodynamic, enzyme-constrained and combined metabolic models has not, to our knowledge, been reported in Aspergillus. Here, we establish a unified experimental and computational platform in Aspergillus niger ATCC1015 by combining curated metabolic reconstruction, Ac/Ds transposon mutagenesis and machine-learning-based essentiality prediction. Our stoichiometric reconstruction, FBA1015, expands reaction coverage and compartmental resolution beyond four published reconstructions. From FBA1015, we developed three models sharing its biochemical network: GECKO light with enzyme-capacity constraints, TFA with thermodynamic constraints and ecTFA combining both. FBA1015 and its derivatives performed competitively in phenotype benchmarks, with added constraints improving agreement with measured growth and gas exchange. Insertion libraries grown in rich medium, glucose or xylose achieved genome-wide densities of one distinct site per 21--112bp across individual libraries. A machine-learning classifier integrated insertion patterns and gene structure to predict essentiality across 11,111 genes. Applying a stringent consensus criterion identified 709 genes predicted essential on both glucose and xylose as sole carbon sources, and 91 predicted essential on one but nonessential on the other. Integration with simulated gene deletions prioritizes candidate genes and testable pathway hypotheses, connecting high-throughput genetic screening with mechanistic analysis for fungal functional investigation and strain engineering.