Monday, 21 July 2008

Answers in Genesis tries to play with the big boys, fails epically.

One of the things that is often used to laugh at how silly Creationists and, more specifically, Intelligent Design, is, is that it makes no predictions, no testable hypotheses, and porponents lack published papers in scientific journals. Some time ago, Answers in Genesis tried to address these criticisms by starting their own journal (Sorry, AiG, but it doesnt count when you need to start your own journal to publish your own papers, and the peer-review process only works when people outside of your own little group review your science). Nevertheless, we all still chuckled because their papers were non-technical drivel with faulty reasoning and laughable conclusions. Now, the folks over at AiG have tried to remedy this by attempting real 'semi-technial' research. In their article, they compare the fitness of an antibiotic resistant strain of Serratia marcescens to a suseptible strain, in an attempt to discredit evolution of antibiotic resistance (and thus, micro- and macroevolution). But, as the folks over at The Panda's Thumb point out, they fail majorly.

The first problem with their research is breaking one of the golden rules of research: if you are examining a particular variable, everything else needs to remain constant. For example, if you are trying to determine if regular or supreme gasoline gives you a better fuel mileage, you have to compare the two types in the same car, or at least the same model of car. You cant put regular in a Civic, put supreme in a Prius and then declare that supreme is better because the Prius goes further. There is no way to determine if the difference was caused by the gasoline itself of if other factors that differ between the two cars were the cause. You'd have to do the experiment in the same (type of) car. Everything except the variable you are testing needs to remain the same. The same thing goes with working with bacteria. If you are going to compare two strains - one resistant to Ampicilin, the other not - then both strains need to be genetically identical sans the resistance gene. This is one of the most basic, common sense rules about research. Its also the first rule that Answers in Genesis breaks in their 'research'. They compared an S. marcescens Amp resistant strain (it's never explicitly stated where they obtained it, but they allude to Dr. Robert Williams at the Texas Medical Center who has aparently kept the strain going) to a "wild type" that they, get this, isolated from a random sample of pond water. They have no way (short of sequencing the entire genome of each strain) if both strains are genetically identical. One is resistant to ampicilin, the other isnt. But is that the only difference? Most likely not. How can they conclude with any certainty that any fitness difference between the two were because of the resistance gene? They can't. It's as simple as that. They cannot make such a conclusion because they used grossly improper research technique; their findings are absolutely worthless.

The second big problem with their research is that they dont understand what the heck "fitness" means. They define fitness as as ‘growth rate and colony “robustness” in minimal media’ (What the heck is a "robust" colony? That's a pretty subjective qualifier). Unfirtunatley, because of their lousy techniques mentioned above, they can't tell if a smaller colony is less fit (by their definition) because of the resistance gene or some other factor they did not controll. Not only that but the figure they use to describe the growth curves has no error bars, so you cannot tell how accurate their numbers are and whether or not the difference between the two strains is statistically significant (it likely isn't). They then go on to conclude that the resistant strain is "less fit" than the wild type. This makes little sense, considering 92% of Serratia marcescens infections in hospitals are antibiotic restistant (which they state in their own paper). If such strains were less fit then why are they more common than the wild type strain? Because hospitals use antibiotics and the resistance provides a reproductive (fitness) advantage. But because they use a lousy definitation of fitness, they conclude that resistant strains are less fit.

The third big problem is that they dont seem to know the difference between "compete" and "compare". What they did in their paper was compare the two strains. They plated one on minimal media and counted the colonies. Then they plated the second on minimal media and counted the colonies. They compared the numbers, etc. Nevertheless, they constantly refer to the bacteria "competing". There is a big difference. Let's imagine we have two cages, each with a mouse. You feed Mouse A ten pieces of cheese, and it eats all of them. You feed Mouse B ten pieces of cheese, and it eats all of them. Comparitively, both mouse seem the same. But put both mice together and feed them some cheese, and mouse A eats 90% of it, while mouse B only gets 10% of it. This is because Mouse A outcompetes Mouse B. This is not what AiG did. They compared the two strains. They didn't make them compete. So how could they conclude that the non-resistant strain of Serratia marcescens outcompetes the resistant strain? They can't.

And even if all this was forgiven, if they had used proper technique, they had made the strains compete, used a proper definition of fitness....their conclusions are still wrong. If they had bothered to look, this exact experiment had been done before (properly), and the authors concluded the exact opposite of what AiG found.

Nice try, Answers in Genesis, but you're not cut out for this kind of thing. Leave the research to the people who know how to do it. Leave the science to the scientists.

No comments: