Hay Fever

compilersex

 

It’s raining DNA outside. It’s also a bright summers day without a cloud in the sky. I am walking through the park on the way to work smelling the intoxicatingly sweet odour of the pollen drifting from the small-leaved lime trees that line the path. Bad news for hay fever sufferers. The pollen count in London is reaching a record high as trees and grasses throughout the Kingdom billow forth their genetic material to the wind.

This fecund explosion of pollen seems so wasteful. Only a tiny amount of this genetic material will end up in its intended place, fertilising the pistils of their flowering counterparts. Moreover, many plants can reproduce asexually by putting out runners, rhizomes and root suckers. Others, like the dandelion, can even produce seeds asexually. When you blow a dandelion ‘clock’, the hundreds of little white feathery parachutes that float on the wind are all genetically identical clones. Unlike pollen, each of these seeds can form a new plant all on its own. This leads many biologists to ponder the question “Why bother with sexual reproduction at all ?”

On the surface, asexual reproduction has a lot going for it. A large number of offspring can be created very rapidly and an exact copy of the genes is passed on which is good news from the individual’s perspective. It takes far less time and energy; no need for elaborate courtship rituals, no peacocks fantail and no deer antlers. But despite these advantages, sexual reproduction is still favoured by nature and the dominant form of propagation on the planet. I have tried explaining to my teenage daughters that sex is an unnecessary waste of time, but they don’t really believe me and still seem to be surrounded by eager looking boys.

The biologist’s answer to why sex is important is sometimes called the ‘Red Queen Hypothesis’. In the late 1970s, W.D.Hamilton demonstrated this with a computer model of artificial life. His simulation started with 200 digital ‘creatures’; some of these reproduced asexually by cloning, others were forced to search for a mate before replicating. These creatures were killed off randomly but, after 100 generations or so, the only ones left were all asexual reproducers. A clear demonstration that cloning beats sex. However, Hamilton then changed the rules and ran the game again, this time introducing viruses that infected the creatures causing them to die. This made a significant difference to the outcome. In fact, the simulation produced a completely opposite result. This time the sexual reproducers won because their greater genetic diversity gave them resistance to viral attack.

In Alice in Wonderland, the Red Queen has to keep running in order to stand still. So Hamilton called his explanation for sex the ‘Red Queen Hypothesis’. He saw life as a constant struggle between organisms and viruses, each one evolving to counteract the other so, in effect, they were running to stand still. Clones, being genetically identical, are highly vulnerable to viruses. In contrast, organisms that reproduce sexually shuffle their genes in each generation which gives them the diversity to withstand viral attack.

If you go for a walk in the countryside this weekend, have a closer look at the hedgerows and you may see another demonstration of this concept. Neat suburban hedges are made of identical plants, bought from the local garden centre and propagated by cuttings. So a suburban hedge is a row of single species clones. But a hedge in the countryside has a wider variety of species; in fact, you can tell the age of a hedge by the number of different species it contains. This is known as Hooper’s Rule. Count the number of species in a 30 meter stretch and multiply by 100 to get an estimate of its age.

 

Hedges

The photo shows the hedge outside my village Church in Chiddingstone. There are five different species in a 30 meter stretch: hawthorne, field maple, yew, oak and blackthorn. This would imply that the hedge is 500 years old, which tallies well with the local historical records. It is also a demonstration that the motto “unity is strength” (see note) is wrong. It should be restated as ‘diversity is strength’ .

The idea that “genetic diversity grants viral resistance” has recently been adopted in a new sphere. Professor Michael Franz at the University of California is looking at ways to introduce a form of genetic diversity into computer application programs. With popular desktop software, like the Windows operating system or the Firefox web browser, every copy installed on a billion of PCs around the world is an identical clone of the master copy. This creates a computing monoculture which is highly vulnerable to attack by viruses, which we have all experienced to a greater or lesser extent. Antivirus software has offered some protection in the past but, as the Economist (24 May 2014) notes, even Symantec (the market leader) recognise that antivirus software is no longer effective and a new approach is needed.

Professor Franz believes the answer lies in changing the way software programs are compiled. Software is written in a high level language, like C++ or Java, but in order to run it needs to be translated into machine code, the binary language that computers understand. This translation process, from high level language to machine code is called compiling. Normally, software engineers want their programs to run as fast as possible so compilers are set to optimise speed. As a crude analogy, it’s bit like asking a translator to convert some text from French to English but to only use a limited English vocabulary with no words longer than five letters.Of course, there are many different ways that a French text could be translated into English; the meaning would be the same but the actual words would be very different once you relax the five letter limit. Likewise, with compiling computer code. Once you are prepared to relax the speed requirement, each compilation can be different from its predecessor. It may run a little slower but it will be a unique instance of that program, different at the binary or “genetic” level from its parent.This ‘multi compiler’ approach has already been tested on the Firefox browser, producing (at least theoretically) a billion different interpretations of the program which are functionally identical but genetically different. When tested against common viruses, they all failed to infect the system and, other than causing the odd crash and reboot, this malware had no effect at all.

This is a promising start and highlights the path to a whole new way of protecting systems from viruses but a commercial implementation is still several years away with plenty of technical issues such as MD5 hashing yet to be resolved. We can also expect that, in time, malware will adapt to this new approach too, as the Red Queen Hypothesis would suggest. In the mean time, the most defence against self-propagating malware is effective and rigorous network monitoring.

We look with hope to the future in anticipation of the multi compiler approach being further developed. There is a neat circularity to this journey. An idea born in Hamilton’s computer simulations of artificial life returns to its roots as a mechanism for defending the computers themselves against viruses. As T.S. Elliot has it in Little Gidding

“We shall not cease from exploration
And the end of all our exploring
Will be to arrive where we first started
And know that place for the first time…”

Footnote : “Unity is strength” is motto that has been used over the years by a number of different nations including Belgium, Haiti, Bulgaria and Georgia. Looking at this list, which are far from the most powerful or coherent countries on the map, is hard to avoid the conclusion that the motto was fundamentally flawed.

Leave a Reply