A SET of techniques developed in California could be the Rosetta Stone that
lets us translate the human genome鈥檚 hieroglyphics. The techniques could show
what the thousands of new genes revealed by the sequencing of the human genome
do, and revolutionise the way we view proteins, the researchers say.
To get a handle on the function of the protein coded for by a newly
discovered gene, biologists look for similar genes whose function is already
known in other organisms. But the new human genome sequence has thrown up
thousands of genes with no known counterparts, leaving researchers in the dark.
While it is possible to discover the function of a completely new protein by
genetic and biochemical experiments, the process is slow and painstaking.
So David Eisenberg and his colleagues at the University of California, Los
Angeles, have devised alternative methods of finding out what mystery genes do.
They use software to study the patterns of proteins across the genomes of as
many different species as possible, from bacteria to humans. The more genomes
you compare, the more you can learn, says Eisenberg.
Advertisement
The first technique, called phylogenetic profiling, involves seeing which
groups of proteins are found together in many different species. For example, if
a group of proteins in a bacterium are also all present in humans, it鈥檚 a pretty
good bet that they all work together in some way.
Another technique, named after the Rosetta Stone that provided the key to
translating hieroglyphics, is based on the observation that the genes for pairs
of proteins that interact in simple organisms, such as yeast, often fuse
together to form a single gene for one multifunctional protein in more complex
animals. By discovering which genes are paired in this way, Eisenberg says he
can predict which proteins might join forces to form big complexes or act
together in biochemical pathways.
Discovering the links between unknown proteins in this way is a step forward,
but the big breakthroughs come if scientists already know the function of one or
two of the proteins in the network. 鈥淚f you know what one of them does, that
gives you a hint,鈥� says Eisenberg.
He has already demonstrated the power of these techniques by applying them to
about 20 fully sequenced genomes. For example, his team was able to assign
possible functions to about 40 per cent of the 1500 unknown proteins in
Mycobacterium tuberculosis, the bug that causes TB, by establishing links
between them and proteins whose function we already know. That meant researchers
could identify about 50 potential targets for drugs鈥攅ssential work since
TB is becoming more and more drug-resistant.
Eisenberg believes the days of studying proteins in isolation are gone.
Instead, scientists will look at proteins in context, as elements in a vast
network of interacting processes.