Phillip Sharp, at MIT, and Richard Roberts, at Cold Spring Harbor Laboratory, independently discover in 1977, both working with adenoviruses, that genes are not continuous segments of DNA as had been universally assumed since the discovery of the genetic code. Using electron microscopy of hybrids formed between mature messenger RNA and the DNA it comes from, both teams observe that the final messenger RNA is considerably shorter than the corresponding DNA region, and that the DNA forms characteristic loops when hybridized with its own messenger RNA — direct visual evidence that intermediate parts of the DNA are not present in the final RNA. The finding reveals that genes are composed of coding segments (exons) interrupted by non-coding segments (introns), which must be removed via a cellular process called splicing before the messenger RNA can be translated into protein. The discovery, completely unexpected and requiring a revision of the fundamental understanding of how genes work in complex organisms (unlike bacteria, where genes are indeed continuous), also explains how a single gene can produce, through different combinations of exons — alternative splicing — several different proteins from a single DNA sequence, enormously multiplying the possible functional diversity of the human genome without requiring a proportionally larger number of genes.