Understanding how proteins and their complex interaction networks convert the genomic information into a dynamic living organism is a fundamental challenge in biological sciences. As an important step towards understanding the systems biology of a complex eukaryote, we cataloged 63% of the predicted Drosophila melanogaster proteome by detecting 9,124 proteins from 498,000 redundant and 72,281 distinct peptide identifications. This unprecedented high proteome coverage for a complex eukaryote was achieved by combining sample diversity, multidimensional biochemical fractionation and analysis-driven experimentation feedback loops, whereby data collection is guided by statistical analysis of prior data. We show that high-quality proteomics data provide crucial information to amend genome annotation and to confirm many predicted gene models. We also present experimentally identified proteotypic peptides matching approximately 50% of D. melanogaster gene models. This library of proteotypic peptides should enable fast, targeted and quantitative proteomic studies to elucidate the systems biology of this model organism.