What had been described by scientists as "junk" DNA is what gives humans their distinct identity.
More than 50 per cent of human DNA are copies of nearly identical sequences and has been referred to as 'junk.' A major source of these repeats is internal viruses that have inserted themselves throughout the genome at various times during evolution.
This research also shows that these repeats are anything but "junk DNA," since they might hold the key to some of the important physical differences that distinguish humans from all other species.
Using the latest sequencing technologies, Genome Institute Singapore (GIS) researchers showed that many transcription factors, master proteins that control the expression of other genes, bind specific repeat elements.
"Our research findings imply that these surveys must also include repeats, as they are likely to be the source of important differences between model organisms and humans," said Guillaume Bourque, GIS senior group leader and co-author of the paper.
"The better our understanding of the particularities of the human genome, the better our understanding will be of diseases and their treatments," he added.
Researchers showed that from 18 to 33 per cent of the binding sites of five key transcription factors with important roles in cancer and stem cell biology are embedded in distinctive repeat families, said a GIS release.
Over eons, these repeats were dispersed within different species, creating new regulatory sites throughout these genomes. Thus, the set of genes controlled by these transcription factors is likely to significantly differ from species to species and may be a major driver for evolution.
"The findings by Dr Bourque and his colleagues at the GIS are very exciting and represent what may be one of the major discoveries in the biology of evolution and gene regulation of the decade," said Raymond White, neurology professor at the University of California, San Francisco and chair of the GIS Scientific Advisory Board.
The paper was published Tuesday in Genome Research.