Devious Distortions: Durston or Myers?
or
Estimating the Probability of Functional Biological Proteins
Kirk Durston, Ph.D. Biophysics
A few days after his 2009 debate with me on the existence of God, well-known atheist, PZ Myers, posted a blog with the title, ‘Durston’s Devious Distortions’, in which he attempted to address my claim that amino acid sequences coding for biological protein families have an astonishingly low probability … so low that they definitively falsify the hypothesis that biological proteins could have been assembled via an evolutionary process.
For his resource material, Myers referred to an online video of a lecture I gave at the University of Edinburgh about a year earlier. That video was for a general, interdisciplinary audience and did not actually discuss how the probabilities of biological proteins are calculated. Given this absence of information, Myers made some assumptions as to how I calculate protein probabilities. It appears that he did not check his own numbers by plugging them back into the relevant equation to see if they gave the results I had obtained. Those who tried to use Myers’ numbers found that they were incorrect, not yielding anything close to the results for SecY or RecA that I had published in the literature and shown in the video. Unfortunately, some atheists who follow Myers, unquestioningly accepted his explanation without checking his numbers, thus spreading Myers’ confusion.
In reality, the probabilities of biological proteins were calculated using a two-step procedure. The first step was to calculate the functional complexity required to code for a particular protein using a method I had published in 2007. That paper was referenced in the video in a slide titled, ‘Case 3: Average protein’ so that inquiring minds could obtain further details. Once a value was obtained, the second step was to plug that value into the equation I showed in the video and solve for the probability. Here, I shall demonstrate how these probabilities can be calculated using real multiple sequence alignment data for protein families.