Changes in DNA are NOT random

Cambridge has a rich history of making discoveries about DNA - the genetic code inside each and every one of us. In the 1950s, Watson and Crick announced that they had unraveled the structure of DNA - the famous double-helix shape. Now, 60 years later, another Cambridge scientist  has made an important DNA discovery. This time about the way the genetic code changes or "mutates" to allow evolution to happen! Bill Amos began by explaining what a mutation was to Graihagh Jackson...

Bill - A mutation is just simply a change to the DNA. If our DNA never changed, you might think that's rather good. But of course, that isn't very good because that means you have no variability. So if a disease comes along, you won't have any variability to fight it, so you can't evolve, you can't change yourself. So you need to balance novelty through mutations but most mutations are actually harmful so you want a little bit but not too much.

Graihagh - When do these mutations happen?

Bill - Well, this is the interesting thing. Previously, until quite recently really, we weren't getting lots of DNA sequence. So we couldn't really quantify mutation rates very accurately and everybody made the perfectly understandable assumption that they just occurred at random like raindrops falling on the road. So, the assumption was that these are so random if you put out a whole load of different buckets in the road, they'd all fill up at the same rate. The really strange thing that my research seems to reveal is that due to the way chromosomes interact. In fact, variability attracts mutations. So this means if you put out a bucket that already had some water in, you'd get more rain falling in it and that's a really strange pattern.

Graihagh - So, what you're saying is, you can now predict where these mutations are going to happen within our genomes.

Bill - The less random the mutation process is, the more you can start to get handles on predicting where these mutations are going to occur. These occurs in a population, but also even in individuals and you may even have applications in looking at where mutations that predisposed cancer are more likely to occur.

Graihagh - So, it could have some really fundamental benefits I suppose to how we screen and predict for things like cancer in the future - medical implications.

Bill - It's really important to understand. I mean, more and more, as we're moving in the genomic era where everybody seems to be collecting DNA sequences from lots of organisms and huge amounts of genetic data from humans, we really are moving into the era where we can quantify very precisely where the variability lies. My research suggests that we can now use this to start predicting more accurately where mutations will occur in the future. And one of the huge benefits in terms of evolution of the sort of phenomena that I've discovered, is that it will attract more mutations towards the genes that really need to be more variable. So, if you're variable to start with, you'll attract more mutations, making you more variable and that helps you direct the mutations towards bits of the genome where they're maximally beneficial.

Graihagh - Going back to your bucket analogy where you're more likely to have all these water collecting in one bucket rather than the other two that suggests that these variations are or these mutations are happening in a cluster. Why are they happening in these hotspots?

Bill - It's the slight freak of how the DNA replication process happens. We've got two copies of each chromosome and when they come together, if there's a point in the chromosome where they differ, for example, you might get a gene from your mother with brown eyes - that codes for brown eyes and one from your father with blue eyes, that's the difference. Actually, the genome will actually recognise  this is a difference. In trying to correct this difference, they tear up a bit of the DNA and relay it. It's like making a little repair job. During the repair, they might make more mistakes. So that's why they tend to occur near to each other.

Graihagh - So, what you're saying is, you can now predict where these mutations are going to happen within our genomes.

Bill - The less random the mutation process is, the more you can start to get handles on predicting where these mutations are going to occur. This occurs in a population, but also even in individuals and you may even have applications in looking at where mutations that predispose to cancer are more likely to occur.

Graihagh - So, it could have some really fundamental benefits I suppose to how we screen and predict for things like cancer in the future - medical implications.

Bill - It's really important to understand. I mean, more and more, as we're moving in the genomic era where everybody seems to be collecting DNA sequences from lots of organisms and huge amounts of genetic data from humans, we really are moving into the era where we can quantify very precisely where the variability lies. My research suggests that we can now use this to start predicting more accurately where mutations will occur in the future.