Probabilities of Abiogenesis and Evolution Debated
Wolfgang G. Gasser
An irrelevant "abiogenesis" paper – 1999-08-01
A huge discrepancy exists in abiogenesis research between what is actually possible and what is generally reported to be possible.
Imagine: the abiotic creation of peptide chains of up to six glycine residues (the simplest amino acid of proteins) under highly artificial conditions (a 100 mM glycine solution dissolved in pure water) is still in 1999 considered such a sensation that this result can be published in Science, one of the two most widely read scientific journals!
Julie Thomas has written an excellent short post on this subject (Elongation of Oligopeptides in a Simulated Submarine Hydrothermal System, 1999, Science , 283).
Julia Thomas pointed out 2 facts which are immediately apparent to an organic chemist.
(1) The largest oligopeptide based on pure glycine was a 6-membered chain.
Why? Well from first semester organic chemistry: one has a competition between intra and intermolecular reactions. Once 4 or more of the monomers are linked together, the amino end can now condense with the carboxylic acid end, forming a ring and that is the end of the polymerization story.
In other words, under even optimized laboratory conditions even remotely relevant polymers won't form.
(2) Why were other amino acids not tested? Well, to break the suspense, here's my hint: why was the ONLY amino acid, glycine, which does not form in a 50:50 racemic mixture, tested? Got it?
Because you cannot force the polymerization to produce pure L-bonds (by excluding the mirror-image and symmetrically identical R-forms present in all the other naturally occurring amino acids). The probability of linking say 400 amino acids, an average-size protein, in only the L-form is 1 chance in 2400. Now include the constraint that only peptide reactions are allowed.
What are the chances the Matsuno "natural" conditions would produce a single correct protein considering only these 2 constraints, given sufficient amino acid? Less than the chance of correctly guessing a particular particle in the whole universe, estimated at around 1 in 1080. (Of course 1 protein alone is not going to go very far in creating or supporting life anyway). We have neglected, of course, the fact that a useful protein requires a specific sequence of different amino acids in the correct order to have any biological activity. In some positions there is simply no leeway if the correct 3-dimensional structure is to form and allow enzymatic activity.
The peptide bonds are thermodynamically unstable, meaning the reversal to starting amino acid is strongly exothermic. Take a wild guess what any water, such as the co-product from the condensation reaction, would do to the amide bond?
It is not possible, with the best of intentions, to see any relevance of this experiment to the abiogenesis discussion.
Abiogenesis virtually impossible (was: Probabilities) – 2000-12-27
The frequentist statistical method, in a nutshell, is this. We calculate the probability that a certain event would occur given a certain hypothesis about the cause of the event. Then, if the probability is very low and we do observe the event, we reject the hypothesis. Obviously, with this method, it would make no sense to calculate the probability of the event given that we observe the event, as that would always equal 1.
The (often unstated) hypotheses given in the case of abiogenesis as understood by modern neo-Darwinians are:
1) no theological concepts
2) no teleological principles or finalistic laws of nature
3) no vital forces or panpsychism
Based on these premises, the probability of a living cell to emerge is indeed virtually zero (in agreement with "notorious folks with a serious background in mathematics").
It makes no sense to exclude all concepts from science which do not agree with the currently prevailing metaphysical premises of it (i.e. with materialistic reductionism), and on the other hand nevertheless stick to a probability of 1 for "scientific abiogenesis".
I think that you need a bigger nutshell. Specifically, you need to say something about rejection regions.
"Rejection regions"? Let us assume you sit in a restaurant and hear the typical sound of coins falling on the floor. You look in the right direction and see around 30 normal coins more or less randomly distributed on the floor.
What will be your reaction if you notice that there are exactly 31 normal coins (with both head and tail) but that all of them show head and none of them show tail?
The probability to normally get either 31 heads or tails is less than 10-9.
Would the fact that the owner of the restaurant is an amateur magician open a "rejection region"?
Let us assume a standard bottle will be found on Mars. How probable is it that a bloody normal bottle could have appeared by chance on Mars?
From the fact that nobody has ever succeeded in building genuine self-replicators, we can conclude that the complexity of the simplest living cells is, at least somehow, higher than of all man-made tools, machines, computers, and so on.
Is it actually reasonable to assume that random self-assembly of a primitive self-replicating organism with the additional property of being able to evolve by random errors into conscious beings is more probable than the random emergence of a standard bottle?
Could our earliest ancestor who represents the transition from abiogenesis to neo-Darwinian evolution actually have been simpler than Adam (Source):
1) Adam was a bacteria-like creature with a cell wall.
2) Adam contained 10 different enzymes consisting of each 100 amino acids.
3) The genome of Adam consisted of 10 x 100 x 3 = 3000 nucleotides.
4) There were 100 copies of each enzyme within Adam.
5) Adam lived in a primordial pond with high proportions of prefabricated amino acids, nucleotides and high energy molecules.
6) These nutrient molecules and waste molecules could easily cross the cell wall.
7) The cell wall protected Adam from harmful molecules.
8) The replication of Adam occurred by
a) doubling the number of each enzyme from 100 to 200,
b) doubling the genome molecule(s) and
c) splitting the cell into two equal parts by randomly distributing enzymes and genome molecules to the parts.
How probable is such a scenario, based on what we have learned from abiogenesis research?
Probabilities – 2000-12-29
But the problem is that the frequentist logic is NOT objective, as advertised. The subjective nature of frequentism is well hidden but it is there.
In my opinion, frequentism is neither more nor less subjective than the alternative approach. Both approaches can be correctly applied in a rather subjective or a rather objective way.
The following extract from a previous post of Richard Wein in this thread shows the main point:
"Let A be the event of abiogenesis (on earth).
Let L be the observation that life exists on earth.
Let H be the hypothesis that the origin of life on earth was purely naturalistic.
Then frequentists want to calculate P(A|H). For frequentists, A is not a hypothesis, it's an event. A frequentist wants to calculate the probability that A will occur, given H. He is not saying that this is the probability that A is true."
Bayesians want to calculate P(A|L). For Bayesians, A is a hypothesis, and they claim to be calculating the probability that A is true. However, they recognize that this probability is a subjective measure, a personal degree of confidence in the hypothesis."
The valid point to which I was referring is that one needs to consider the fact that life exists on Earth. Probabilities computed without that condition are not relevant.
Probabilities computed based on transparent hypotheses are relevant even if they are far away from what is observed in reality. They are evidence for or against the concerning hypotheses.
Based on our theories about physics and chemistry it is at least in principle possible to make the synthetic apriori judgement (Immanuel Kant) whether abiogenesis as conceived by neo-Darwinians is possible or not.
You wrote in the thread "ID and the anthropic principle":
I read both Dembski and Behe to be asserting that if the probability is small enough, then it is literally impossible for the outcome to occur. That is, Dembski says that if the probability of a specified event is less than 10-500, then it will not occur at all (i.e., is impossible). This flies in the face of probability theory, but it seems that this is what he is saying.
Do you seriously suggest that an event with a probability of less than 10-500 is possible in a finite universe of less than 10100 galaxies and an age of less than 10100 years? I agree with Dembski and Behe that such an event is virtually impossible.
I'm convinced that simple probability estimates do not only refute both abiogenesis and neo-Darwinian evolution but even life itself. This can be shown by applying the frequentist approach using the hypothesis that the currently accepted chemical and physical laws (e.g. Brownian motion) are essentially correct [and that nothing outside these laws is involved].
One single prerequisite of a successful cell division is the condition that enough enzymes find their corresponding substrates in time. That cell-replication actually happens in nature must not be confused with the question, whether it is consistent with the currently prevailing theories of physics and chemistry.
I discussed such issues March 1999. Here some extracts:
It is necessary to have a concrete imagination of the proportions between cells, enzymes, molecules and so on. Therefore I have introduced the enlarged model where 1 mm corresponds to 1 nm. The 'diameters' of enzymes are then in the order of a few millimeters and the 'diameter' of a water molecule is about 0.3 mm (there is room for 33.3 water molecules in 1 cubic millimeter).
Many enzymes work at defined places in a cell. If we create such an enlarged model where enzymes are like little balls of a few millimeters, then the volume of the whole cell is about 1000 cubic meter. Imagine concretely this situation:
A little ball must come very close to a substrate and substrate recognition even depends on the correct alignment of the little ball. In addition to that, enzymes often have to pass cell membranes in order to reach their destination. What is the moving force of enzymes? It cannot be electromagnetic attraction or repulsion. So the moving force must primarily depend on random thermal motions (as Brownian movements do).
Please imagine very concretely the many water molecules (0.3 millimeter) colliding with the enzyme (e.g. a diameter of some millimeters).
talk.origins probability-abiogenesis FAQ criticized – 2001-01-04
Extracts from the FAQ of talkorigins.org:
Problems with the creationist "it's so improbable" calculations:
The genuine problems are rather on the neo-Darwinian side. By the way, I'm an uncompromising evolutionist (i.e. I'm convinced of a continuous emergence of the world and of life).
1) They calculate the probability of the formation of a "modern" protein, or even a complete bacterium with all "modern" proteins, by random events. This is not the abiogenesis theory at all.
A probability of 10-100 is enough to refute abiogenesis, so it doesn't help to criticize calculations resulting in e.g. 10-1000000.
For comparison: in the order of 1020 milliseconds have passed since the birth of the earth, a planet consisting of in the order of 1050 atoms. This results in 1070 atom-milliseconds. It is obvious that the number of any reasonably defined "abiogenesis events" is by many orders of magnitude lower than this number of atom-milliseconds.
Even if we assume that 1) around 1020 planets could have given rise to us, and that 2) one single successful abiogenesis event leads inevitably to higher forms of life, a probability of 10-100 would still make our existence very improbable.
Because neo-Darwinism is not even consistent with the evolution of the upright gait in humans, assumption 2) is obviously untenable.
2) They assume that there is a fixed number of proteins, with fixed sequences for each protein, that are required for life.
This assumption is not necessary in order to refute (purely materialistic) abiogenesis, because the improbability lies already in functional constraints which are a prerequisite for any form of self-replication.
If six proto-enzymes with each a probability of at most 10-10 are necessary for self-replication to start, we get already a probability lower than 10-60.
The probability of random emergence of a proto-enzyme cannot be higher than 10-10 even if it consists of only 10 rather widespread building blocks.
Such an emergence probability must include both the probability that a correctly bonded system of 10 building blocks arises and the combinatorial probability. For instance (0.5 x 0.2)10 results in 10-10, where 0.5 is the probability of correct bonds per building block and 0.2 the combinatorial probability per building block.
The FAQ ("Coin tossing for beginners and macromolecular assembly") completely ignores the question of the bonding probability. Only the combinatorial probability of a (32 amino acid long) sequence is taken into account and the experimental fact that correctly chained 32 amino-acid long sequences do not arise abiotically is discarded.
And we must keep in mind that the final product of abiogenesis must be able to undergo neo-Darwinian evolution and to survive a change in its environment (e.g. drying up of the primordial pond).
If "not even a protobacteria" is necessary for further evolution by replication with mutation and selection to start, "but one or more simple molecules probably not more than 30-40 subunits long", then we still remain with the fact that the average probability of correct bonds with viable neighbors is certainly lower than 10% per subunit. So the probability of the emergence of one "simple molecule" of this kind is still lower than 10-30.
The abiotic emergence of RNA enzymes is even more questionable because it contains ribose. Ribose is not even resistant to water of normal temperature (see -> ribose).
"Like many monosaccharides, ribose occurs in water as the linear form H-(C=O)-(CHOH)4-H and any of two ring forms: ribofuranose ("C3'-endo"), with a five-membered ring, and ribopyranose ("C2'-endo"), with a six-membered ring. The ribofuranose form is predominant in aqueous solution." (Source)
3) They calculate the probability of sequential trials, rather than simultaneous trials.
This statement can be interpreted either as a strawman or as a summary of a central error of the FAQ, namely the assumption that whole proto-enzymes emerge by simultaneous trials, and not by sequential steps (i.e. addition and breaking off of building blocks). (See -> 21-Mar-1999)
4) They misunderstand what is meant by a probability calculation.
Those who declare simple and logically correct common-sense reasonings wrong in order to save the currently prevailing "scientific" dogmas are the ones who "misunderstand what is meant by a probability calculation".
5) They seriously underestimate the number of functional enzymes/ribozymes present in a group of random sequences.
The low actual number of functionally equivalent proteins and even DNA-coding sequences is rather evidence against the relevance of this argument (See -> The Death of Neo-Darwinism).
The fact that each of e.g. 90% of the amino acids of a protein can be replaced by others without loss of function does not entail that the protein would still work if most of these replaceable amino acids were replaced at the same time.
But even if we accept the argument, in this context it does not change a lot whether the combinatorial probability of a 100 amino-acid long enzyme is 20-100 (with only one viable variant) or (20/5)-100 = 4-100 (with 5100 viable variants).
The formation of biological polymers from monomers is a function of the laws of chemistry and biochemistry, and these are decidedly not random.
The laws of chemistry and biochemistry do indeed lead to the needed chemical bonds. Yet these laws also lead to undesired bonds and to the decay of desired bonds. If peptides are synthesized in vitro, special methods are needed in order to prevent undesired chemical bonds.
However these laws do not concern the question whether a final sequence is able to fold into an enzyme and to carry out complex tasks. So these laws have rather to do with the probability of correctly chained sequences long enough to arise at all than with the combinatorial probability of such sequences.
talk.origins probability-abiogenesis FAQ criticized – 2001-01-11
cannot reply to your criticism of
my critique of your probability-abiogenesis FAQ without repeating myself. So I
only refer once again to my previous writings, especially from March 1999.
Only one short remark: even if the prebiotic ocean contained once an amino acid concentration of 10-6 M, it certainly would have consisted of a racemic mixture of R-forms and L-forms.
By the way, does anybody know an experiment showing relevant enzymatic activity of proteinoids (surpassing catalytic activity of atoms and simply molecules), consisting either of a racemic mixture or exclusively of mirror-form building blocks? In the case of abiogenesis research, experiments with mirror forms should be taken more seriously, because we can be rather sure that it is not primarily the work of preexisting enzymes or ribozymes which is responsible for the results.