On Being Wrong in Science | The BioLogos Forum

John Walton’s post on the BioLogos blog titled “On Being Right and Wrong” takes a nuanced approach to questions of “right” and “wrong” in discussions about science and faith. Reading it brought to mind a time in my biochemistry lab when the right result came from a wrong beginning.

With complex theories, right and wrong have a complex, even symbiotic, relationship. If the right science can start with a wrong idea, then scientists can benefit from being wrong. At least, I did.

Soon after starting my own research lab a decade ago, I had a great hypothesis and was ready to prove it right. I work with proteins, and I had the idea that my proteins would work better if we tightened them up and made them more stable. Proteins are like tiny balls of yarn, so we changed their internal chemistry to wind them more tightly.

Our first tightly-wound protein worked no better than normal proteins. A second protein gave the same result. After a dozen proteins failed the test, I began to suspect that my hypothesis might be a bit flawed. At that moment, I felt a tension in my gut, a particular form of anxiety experienced by all scientists. The pressure of “publish or perish” became tangible.

But then, thankfully, something went wrong with my wrongly directed experiments. These two wrongs made a right. By accident, one of our proteins had a mutation that made it less tightly wound. This protein worked better in our tests. So we changed direction and made more loosened proteins. Most of those also worked better. After making two dozen of these, we were able to publish the results, and my lab did not perish.

I never would have tried loosened proteins at first, but I had to go through the simple, wrong hypothesis before finding that reality was more interesting. The wrong idea led to a right result. Some days, being a scientist is more about fixing where you’re wrong than it is about being right in the first place.

This happened for me with the theory of evolution as well. In high school and university, I adopted arguments against evolution that I now believe are wrong. You may recognize some of them by familiar phrases that often accompany them: “simple organisms have more chromosomes than complex ones,” or “the flagellum appears irreducibly complex,” or “some scientists once labelled some DNA as junk but were proved wrong.”

Today, I think these arguments are wrong, even if some start with technically accurate statements. These are inadequate to sufficiently outweigh the thousands of other experiments that support the validity of evolutionary theory. It took decades for me to consider and accept the counter-arguments, because there are so many, but gradually, I agreed with the scientific consensus, that evolution is the right explanation for life on this planet.

Because the process happened over decades, with each new experiment I was able to see how my faith still fit. I was also able to see the beauty and elegance in each experiment. In the end, my faith in God as Creator was not displaced, but enhanced, by understanding natural history through evolution. I see God’s faithfulness in the billion-year regularity of the natural laws that allowed life to grow, adapt, and organize into people that can choose to worship.

In the edifice of science, the anti-evolution arguments above are not what I would call load-bearing arguments—arguments that, proven valid, would single-handedly disprove evolution. Even if these arguments are true and their contribution knocks out a section of the edifice, the structure still stands because the network of evidence is so much larger than what the argument at hand addresses. It’s like removing a single Jenga piece from a Jenga stack the size of the Great Pyramid—very rarely does a single issue take out a structure that substantial.

The genetic evidence that is consistently accumulating in academic journals (some of which is explained by Dennis Venema on the BioLogos blog) continues to convince me that the theory is right. I was therefore wrong to attack it, and, like with my protein experiments, I have now changed direction and adopted a new theory with new hypotheses.

These new hypotheses led me to see a chemical order undergirding evolutionary change. In order to explain this chemical order, I received a BioLogos Evolution and Christian Faith grant to write a book titled A World From Dust: How the Periodic Table Shaped Life, to be published this winter by Oxford University Press.

This narrative interpretation of natural history comes from the fact that we have been given a universe in which science is possible—a universe of constancy, which is consistent and comprehensible on a billion-year scale, filled with regular events that we can understand, replicate, and predict.

From the proper perspective, chemistry and math can even predict some of the paths of evolution. Those paths can be complex. At times, life endured intense threat or meandered through swamps of random events, but even then, the paths were shaped by consistent chemical rules.

These rules emerged from the organization of the periodic table, resulting in a chemical sequence running through natural history. A World From Dust describes how this orderly sequence was first predicted by the chemist R.J.P. Williams, and later supported by genetic data.

Real stories follow complicated but comprehensible trajectories, starting with something wrong and turning it into something right. In the greatest of these, Adam’s sin marred God’s good creation but led to Christ’s redemption—the “felix culpa” of Easter morning. Scientific stories can follow the same pattern if, as scientists, we listen carefully to the world God has made, willing to change when we’re wrong. What is wrong can be made right in a recurring story of hope.


This is a companion discussion topic for the original entry at https://biologos.org/blog/on-being-wrong-in-science

Please post here if you have questions about any of this – whether it’s the science of protein design, the narrative of natural history, or the theology of my own journey. More information on my upcoming book is at benmcfarland.com.

(For more on the ideas in the next-to-last paragraph, see RJP Williams’s The Chemistry of Evolution; for more on the ideas in the last paragraph, you might start with CS Lewis’s friend Owen Barfield’s book titled Saving the Appearances.)

dear dr ben. you said that: "thousands of other experiments that support the validity of evolutionary theory. "

from what we can tell there is no even one experiment that support the claim that an eyespot or a flagellum can evolve step wise from non eyespot\flagellum. so there is no even one experiment that support the central claim of the evolution theory.

you may say that it need milions of years. but then it make it only a belief. a tipical protein is about 250-300 amino acid long. experiment like this show us that for a functional protein about 150 aa long we will need near 10^70 mutations:

so even 4.5 bilion years its nothing.

now you may say that maybe this protein evolve from a simpler one. but once again experiment show us that a big part of the protein size need for its minimal function:

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC211289/

again- experiment support the creation model.

have a nice day

Hi Dr. McFarland, @benmc,

I’m curious what you consider the “edifice” of evolution to be. In his book, The Edge of Evolution, Dr. Behe describes the edifice of Darwinian evolution to be Natural Selection, Common Descent, and Random Mutation. Behe accepts and even argues for both Natural Selection and Common Descent. But he also argues that not all mutations were random, and that many had to be non-random. He then suggests that these non-random events were designed, perhaps being part of the Universe that was chosen to be actualized from a vast array of possible universes. It sounds like Behe might accept much of what you write, but that he would suggest that there is still a need for many events that were neither the result of chance or law.

I think your analogy of the Great-Pyramidic Jenga stack is also applicable to foundation of scripture. For me personally, accepting the scientific arguments in favor of evolution was not much of a struggle; the real issue for me came down to what the acceptance of those arguments did to my interpretation of the Bible as a result.

At first I thought (and I think much of evangelical Christian culture states directly) that removing or relocating any one block on our biblical Jenga stack (e.g. a strictly literalistic reading of Genesis 1-3) will cause the entire thing to collapse. Realizing that that Jenga stack is enormous and enduring enough to support my doubts, questions, and reinterpretations was what ultimately allowed me to accept the fact of evolution alongside my faith in God.

(Shout out to BioLogos for being a HUGE part of that journey.)

3 Likes

Good Blog.

One thing science and faith have in common is that we need to learn from our mistakes.

Sin is a mistake. If we do not learn from it and repent or change, we are lost. Sadly too many people are more intent to use theology to prove they are right, rather than learn from their mistakes.

The truth is we never know everything. We can always learn more. Jesus taught us that our faults are not our mistakes and failures, but our failure to learn from our failures.

When science and faith do not agree, there are three possibilities. 1) the Science explanation is wrong, 2) our understanding of our faith is wrong, and 3) both are wrong.

For evolution, the research has concentrated on the genetic basis of variation, which is basically sound. On the other hand no one has been able to scientifically verify the Darwinian view of survival of the fittest, based on Malthusian population theories which have been widely dismissed, even though persons as diverse as Lynn Margulis and Karl Popper have pointed out that the process of natural selection has not been scientifically established.

Science is nothing if it is not based on experimentation or closely defined field studies. Science fails when it fails to follow its own guidelines and test its own assumptions.

Evolution is not wrong, but neither is it thoroughly understood. Some scientists dissatisfied with the current state of the science have come together as the Third Way. There is no reason why other people should not make responsible criticism of evolution, even though ID does not seem to me to be helpful.

Hi Ben

I sure wish we could get together for an evening of discussions, dinner, and perhaps a glass of wine! I would like to learn more of your proteins–tightly or loosely wound. Together with Prof. Corwin Hansch, I helped design a program that would calculate from structure the hydrophobicity of any molecule. As you well know, hydrophobicity and hydrogen bonding are the main solvation forces that determine how a protein folds. The program does quite well for small drug molecules. (It is in use by Pfizer, AstraZeneca, Glaxo, Abbotts etc.) But when we applied it to peptides, we got as far a penta- and things got too complex. Mother Nature was patient and took a billion years; we gave up after a few months.

As you can already tell, explaining all the nuances of evolution to folks who have been brought up on a literal reading of Genesis takes patience and care. I want to read your book as soon as it has been published. My niece, Diane Sweeney, also received a BioLogos grant, and I served as a kind of ‘scientific advisor’ for the video she and Josh Hayashi produced: “Author of Life”. It was severely criticized by Ken Ham, but I took that as a complement.
Al Leo

2 Likes

I’ve thought a lot about those experiments over the years. From working with proteins for a while, I’m pretty sure that there has been more than enough time for proteins to evolve. It would take a long time to list all the reasons, but as just one example, metal ions can catalyze certain reactions all by themselves, which requires no protein assembly. If the proteins bind to themselves (and/or to metals), they can make bigger structures from just a few amino acids (this is how many anti-microbial peptides work for example) – so you don’t need 150 amino acids. Origin of life is a fascinating problem and a very high bar to get over, but in my book I have a chapter about how even that might be becoming clearer with experiments in the lab that more closely approximate the early Earth. And I also think that a model in which the universe is gifted to produce life from chemistry could also be considered a “creation model”! There’s not really room here to support my conclusion since it took so long, but some of this is in my book, too (but that will take at least a few more months to come out). Yours, Ben

1 Like

I don’t think all mutations were random either – but I think the major category of non-random mutations would be stress-induced mutations and large-scale genetic movements like whole genome duplication. And these “non-random” classes of “mutations” and the non-random nature of chemistry constrained the “random” movement of atoms and point mutations (lots of quotes there showing how much science this is summarizing in a sentence!). The chemistry of our universe is indeed special, although I would not go as far as Behe in saying it would compel someone to say it was designed – I think it’s consonant with a grand narrative and natural theology like Alister McGrath suggests. Behe’s notion of design is too mechanical and “solid” whereas what I see of life’s design is much more fluid and contingent – more life-like. I also think that natural selection when constrained by chemical laws is capable of building complexity and recording information. I think Behe would disagree from what I can tell from his Edge of Evolution book. Lots of ideas behind these words, too many for a comment! Basically, in the mid-90’s I was fully on board with Behe’s ideas but since then I haven’t found them to be fruitful or to work for me as a scientist. RJP Williams’s ideas are different, for example – I can build on and with those ideas, and that’s where my book came from. Thanks for reading. Yours, Ben

1 Like

I’m trying to find that third way as well. I think RJP Williams, Eric Chaisson, Simon Conway Morris, and Adrian Bejan (all in my book) may offer a way forward (and just read a great book by Terrence Deacon that has just joined that list). This means that some prominent scientists are wrong (in the complex and qualified way I used that word in the post!). In my book I single out Stephen Jay Gould’s book Wonderful Life because I think it’s particularly and demonstrably wrong (or at least incomplete). Hope to continue this conversation when the book’s out! Yours, Ben

1 Like

Hi Dr. McFarland @benmc,

I think Behe would point to the recent work of Joe Thornton’s group, which showed that the evolution of a hormone receptor protein was extremely unlikely, and either the result of historical contingency or design. Since this was the first protein whose evolutionary steps had been worked out, Behe argues that it’s reasonable to expect that many other proteins are equally the result of either historical contingency or design. Unless others have done a lot of similar work to Thornton’s since then, that show the opposite, it sounds like your conclusions may be premature, if not mistaken.

This gets to the core of one of the things I think science has shown about our universe. Thornton’s work is great and has a place of prominence in my book (Chapter 12 to be precise). I’ve been following it since he presented it at a conference before publication a few years ago.

I interpret it as showing very clearly the concept of levels – that at lower levels, you can have random or very historically contingent effects, while upper levels are more predictable. In this case, the level is that of protein shape, which is historically contingent in this case. The protein shape is contingent and the exact shape of the hormone is also contingent – but the composition of the hormone (oxygens arranged around a four-carbon-ring center) and the need a hormone to communicate among organs, those are predictable.

The higher up you go in the levels of life the more predictable and less historically contingent life gets. Biologically, species are pretty much contingent, but ecosystems and “chemotypes” (RJP Williams’ word) are predictable. At the lower levels, atoms, nucleotides, amino acids, and exact shapes may flow unpredictably like water molecules in a stream, but when you step back to look at the stream you see predictable patterns over time – which can be told as stories.

Here’s another article that backs up what Thornton said but on a wider scale (very specifically at the protein sequence level): http://www.pnas.org/content/112/25/E3226.abstract

A week ago I talked about these levels on my blog and cited a good article that seems to agree with this concept: Arrow Through the Sun: A World From Dust (Plus): The Higher the Level, the More Repeatable the Evolution

So, yeah, Thornton’s contingency is real but I think it’s built from predictable chemistry (oxygen + sterol = hormone) and produces predictable outcomes at the higher levels (chemotype/ecosystem/planet). Yours, Ben

Great story about the hydrophobicity program! Yes, I know Diane from the ECF Workshops, and “Author of Life” is great work. We’re working on reengineering the proteins now by linking them in new ways, although that’s been on hold while I’ve been writing. Some new publications have come out since the one I linked in the caption above but there’s a few years of unpublished work still to come. Yours, Ben

@benmc,

Well you’ve certainly piqued my curiosity. I look forward to reading your book.

1 Like

@benmc

The chemistry of bio-systems is truly fascinating and while I would disagree with your terminology of ‘fluid’ I agree that there is such an array of factors that it is difficult to discuss mechanisms such as chemist may propose, as for example, catalytic chemical reactions. The bio-systems are influenced by hydrophobic/hydrophilic interactions, hydrogen bonds, 3D structures/configurations, stereochemistry, changes in relative concentrations that may impact of very large molecules – and I do not think this is at all a complete list.

Consequently I find a non-random (or one which seeks constrains in bio-systems) overall outlook appealing when considering the enormous complexity of the bio-world. This outlook may form part of an scientific project, whereas design by an external agent seems too far removed from the scientific enterprise proper. I guess this outlook would fail to please neo-Darwin purists, and also fail to please those who seek to find an intelligence directing the bio-world.

I am not sure I understand what you mean by predictable patterns in your statement, “At the lower levels, atoms, nucleotides, amino acids, and exact shapes may flow unpredictably like water molecules in a stream, but when you step back to look at the stream you see predictable patterns over time – which can be told as stories.” I would think the ‘high’ level of biology offers less predictability as scientists understand it, whereas the chemistry of atoms and molecules is an exact science and predictable.

I agree (assuming you’re talking about species as the basic “unit” of biology)! Let me try to clarify: The levels I’m talking about are more within the disciplines than between them or identified with specific disciplines. (Maybe I’m reluctant to consider my beloved chemistry “lower” than biology!) Ecosystem structure is more predictable than the characteristics of a specific species, within biology. The pressure of a gas is more predictable than the random movement of single atoms withinchemistry (or, say, the predictable shape of the Boltzmann distribution emerges out of randomness). The ordered columns of the periodic table emerge from the randomness of quantum mechanics and the orderliness of math.

RJP Williams has this concept of “chemotypes” which are groups of species assigned by chemical function. Photosynthesizing organisms would be one chemotype, oxygen-using animals another. The predictability in chemical sequence that Williams talks about applies to these groups of organisms much more than to individual organisms. This is what I hope to talk about! Hopefully my book can make this a little more clear because it’s a major theme. Yours, Ben

hi de ben. yes, catalytic reactions can be done without any enzyme. but to evolve a protein that make those reactions we will need a big protein. this is because several reasons:

1)for example: to make a protein that bind 2 substrates we will need at least 2 binding sites. so even if this reaction can be done without this enzyme, the transition from reaction without enzyme to the enzyme need a lots of amino acid.

2)a lots of proteins are structual proteins. so they not have any conection to catalytic reactions.

3)a lot of reactions need a multi steps reactions. so they cant made wihout several enzymes.

by the way, im not talking about origin of life but evolution. and according to the evolution a lot of new proteins evolve from another proteins. so i show that the chance to this claim is very low.

have a nice day

I work with proteins, and I had the idea that my proteins would work better if we tightened them up and made them more stable.

Ben, I recognize that your description was simplified for a non-technical audience but you might want to expand on the hypothesis a bit. “Better” in what regard? (e.g. Operation at higher temperatures or in solvents? Faster turnover? Longer lived?) That hypothesis can’t apply generically as work on protein dynamics over several decades has demonstrated that it’s not an either/or situation. You must have had a very specific enzyme or application in mind. Would you describe it because it sounds interesting.

As a general comment: Just about every experiment that doesn’t quite work as it was expected tends to spawn and/or kill any number of hypothesis about what just happened. But if you use a checklist when running the experiment you greatly reduce the odds you’ll have to investigate a trivial hypothesis like, “I forgot to add the MgCl2 to the buffer”. :smile:

1 Like

Glad someone wants to know about the biochem! We were working on an immunoreceptor-protein interaction and were trying to make it last longer/stabilize it. The disordered region looked like it was getting in the way of that and we had a protein design algorithm to introduce more order. But it wasn’t getting in the way! The protein was stabilized (we checked by CD) but the protein-protein interaction was slowed. All the details can be found in the JBC study linked in the caption.

And yes, there’s lots more ways to be wrong than there is to be right, and the trivial causes outnumber the significant. Brings to mind a quote that some artist said to his critics: “Obviously, it is easier to kill than to create.” !! Controls are required.

Yours, Ben

This topic was automatically closed 4 days after the last reply. New replies are no longer allowed.