Our aim was to understand mechanisms for clustering and cross-linking of gliadins, a wheat seed storage protein type, monomeric in native state, but incorporated in network while processed. The mechanisms were studied utilizing spectroscopy and high-performance liquid chromatography on a gliadin-rich fraction, in vitro produced alpha-gliadins, and synthetic gliadin peptides, and by coarse-grained modelling, Monte Carlo simulations and prediction algorithms. In solution, gliadins with alpha-helix structures (dip at 205 nm in CD) were primarily present as monomeric molecules and clusters of gliadins (peaks at 650- and 700-s on SE-HPLC). At drying, large polymers (Rg 90.3 nm by DLS) were formed and 13-sheets increased (14% by FTIR). Trained algorithms predicted aggregation areas at amino acids 115-140, 150-179, and 250-268, and induction of liquid-liquid phase separation at P- and Poly-Q-sequences (Score = 1). Simulations showed that gliadins formed polymers by tail-to-tail or a hydrophobic core (Kratky plots and Ree = 35 and 60 for C- and N-terminal). Thus, the N-terminal formed clusters while the C-terminal formed aggregates by disulphide and lanthionine bonds, with favoured hydrophobic clustering of similar/exact peptide sections (synthetic peptide mixtures on SE-HPLC). Mechanisms of clustering and cross-linking of the gliadins presented here, contribute ability to tailor processing results, using these proteins.