November 21st, 2003


Sydney Brenner

We had a seminar today at work given by Sydney Brenner. He's one of the Grand Old Men of molecular biology so this was a very well attended talk - there were people sitting in the aisles in the lecture theatre, which is forbidden by fire regulations so they had to go outside and listen on a video linkup, and this is a fairly large lecture theatre (I think it seats two or three hundred).

Sydney Brenner is famous both for discovering the Central Dogma of molecular biology, and for setting up the study of the model organism C. elegans. The Central Dogma of molecular biology is that DNA makes RNA makes protein. To a modern biologist this is just one of the obvious facts of life, but back in the early fifties when Watson and Crick had only just elucidated the structure of the DNA molecule it was completely unclear how the information encoded in the DNA helix (by the order of the four bases (A, T, C and G) a quarternery equivalent of a computer's binary encoding) was read and acted on by the cell. Brenner discovered the molecules called messenger RNA (mRNA), and showed that a gene (a stretch of DNA) is transcribed as an mRNA molecule. This is then translated by a complex called a ribozyme which builds the protein molecule by stringing together amino acids following the instructions given by the RNA. He also was able to show that it is non-overlapping three letter groups of bases of the RNA that specify each amino acid - there are 64 possible triplets, and the 20 amino acids and the stop code are specified by multiple triplets, generally but not always with the third base in the triplet being redundant.

After this great discovery he turned his attention to the (still unanswered) question of how this production of proteins creates an organism, he was instrumental in getting funding for and setting up the research into C. elegans. This organism is a microscopic worm, with a very short generation time. One of its other advantages in the study of development is that at every stage each cell gives rise to exactly the same daughter cells and structures in every worm of the species (which is not the case in vertebrates), so it is possible to study the effect of particular mutations on cell lineage and differentiation more easily.

His talk today was interesting in two ways. Firstly it was a very old-fashioned talk (he was born in the twenties after all) - you don't realise how much you expect visual aids until someone gives a talk without them. There were no slides, he stood and talked to us. In fact it didn't even look like there were any notes. I'm not sure if slides would have distracted from what he was saying, or if they would have been a useful aid for concentration.

It was also (of course) interesting for the content - it was a very general talk on that state of the field as he sees it, and where he thinks it should go for the future. Some of it really struck a chord, we'd been talking in the lab the other day about the proliferation of -omics in the field, someone had started talking about "degradomics" which they defined as the study of all the proteinases in repertoire of cells. It's the current fashionable thing to be, a bit like the proliferation of -ologies in the past. But, Brenner was asking, is it really the most relevant or necessary thing to be studying? Surely we shouldn't be as interested in just collecting large amounts of data on something like the proteinases that are present in a cell, but in more functional groupings of information. His analogy (paraphrased by me) was that of a factory line where you take in the raw material and it undergoes a series of transformations with product coming out the other end. Some people study the processes that occur within the factory, some study the presence or absence or involvement of various factories in the cell. But is it really a relevant thing to do to be saying "here's some scissors, here's a knife - they both cut, this factory has one, this has the other and this one has both"? That doesn't actually tell us anything real about any of the factories. And Brenner's view is that a large subset of bioinformatics is at that sort of level. So the main thrust of his talk was that we need to move beyond this cataloguing of observations and go back to asking proper questions with our experiments.

There were other things that I was less sure about - for instance I was less convinced by his insistence that all databases should be Complete, Accurate and Permanent to be useful. Whilst that's the goal to strive towards I feel that it is also useful for the incomplete, inaccurate and temporary databases to be released into the public domain for others to contribute to and use.

But that is a minor quibble about a relatively small part of the talk. It wasa great experience to listen to one of the founders of the field I work in talk about his ideas and his vision for the future of the science.