A gene in a complex organism is one continuous stretch of DNA that directly matches its messenger RNA.

The early molecular picture of a gene was very direct. DNA was copied into messenger RNA, which carries genetic instructions that can be used to make a protein. It seemed natural to expect the finished RNA to match one continuous stretch of the gene that produced it.

In 1977, Richard Roberts and Phillip Sharp independently found something different. When they compared messenger RNA with the DNA it came from, the RNA matched several separated pieces of DNA. Stretches in between were missing from the finished message.

The first RNA copy is therefore better thought of as a rough draft. It can contain sections that will remain in the finished messenger RNA, called exons, separated by sections that will be removed, called introns. The cell cuts out the introns and joins the exons together in a process called RNA splicing.

This also made another possibility clear: the same gene can sometimes be spliced in different ways, keeping different combinations of exons and producing different RNA products.

Many bacterial genes are continuous, and some genes in organisms with cell nuclei contain no introns. What failed was the rule that every gene must correspond to one uninterrupted stretch of its finished messenger RNA.