Chance, design and artificial life

In previous posts in this blog I have mentioned my experiments on artificial life: the simulation in a computer of processes similar to those that take place in living beings. Artificial life should not be confused with synthetic life: construction of artificial living beings in the laboratory.

One of the most used tools in artificial life (and in other related fields) are genetic algorithms, which simulate biological evolution within the computer, and make it act on the entities that are the subjects of the research. In these experiments, a mixture of chance and necessity (the title of Monod’s book mentioned in the previous post) is used. Chance is usually applied with a pseudo-random number generator that modifies the operation of the rest of the algorithm, which represents necessity.

Intelligent design or random evolution?

Charles Darwin

As all scientific theories, the theory of evolution will always be provisional, but in a century and a half it has been quite well contrasted. It's not likely that a scientific revolution will declare it wrong or obsolete, although perhaps there will be fine tuning, as with Newton's physics and Einstein's theory of general relativity. An attack on the theory of evolution should be based on finding discrepant facts, which up to now have never appeared.

The problem is, some of those who defend the theory of evolution go one step further and offer philosophical speculations and dogmatic statements as though they were testable scientific theories.

As every scientific theory, the theory of evolution is a set of hypotheses that try to explain certain facts. It is based on the verification that species change, and studies the mechanisms that can lead to this change: mutations, DNA, natural selection... Added philosophical connotations are not scientific, whether it is affirmed, with believers, that there is intelligent design; or with atheists, that everything is a consequence of chance.

Ant colonies, real and virtual

Formica fusca

Ants, hymenoptera related to wasps, are social insects. An anthill or ant colony can contain, from a few dozen individuals, to over half a million. The number of castes varies, depending on the species, between three (fertile males and females, sterile female workers) and over twenty. The feeding a larva receives decides the caste to which it will belong.

Strange forms of parasitism have arisen among ants, as in Amazons ants (Polyergus), whose workers specialize in fighting and starve in the presence of food, unless a worker of Formica fusca feeds them. To seize these auxiliaries, the Amazons attack the nests of Formica fusca, kill their queen and enslave the workers. In extreme cases, such as ants of the Anergates genus, the queen invades a nest of Tetramorium, supplants its queen, and fed by the workers of the other species, produces eggs that become queens and males, but no workers, which are not needed.

Evolution in social insects probably reached the highest levels of instinctive complexity that can be achieved with a nervous system as limited as that of arthropods. In the tens of million years since the origin of these societies, evolution has introduced secondary changes, which have led to great diversity: there are more than three thousand species of ants, but there seems to have been no progress in their social structure. They are highly successful animals, very abundant, and spread throughout the world, but stagnant.

What is artificial life?

Thomas S. Ray

As I said in an earlier post, artificial life is a branch of computer engineering that builds programs that emulate the behavior of living beings: artificial living beings, or colonies of living beings, such as anthills or hives. Since I have worked in this field, I’ll tell here a little about artificial life.

In 1991, Thomas S. Ray built a program he called Tierra, where a series of artificial organisms evolved and competed for the available resources in the computer. These resources were essentially the computer memory, which was limited, and execution time. The objective of each individual was to copy itself into a piece of available memory. When copied, however, errors (mutations) could be introduced, so that the organisms in question were able to evolve.

The execution took place in a virtual machine equipped with a simple machine language, with 32 different instructions. The individuals were programs made of instructions written in the machine language. Some basic instructions were relatively complex, such as asking the operating system to allocate a certain space. Although very simple, the original program was able to copy itself (with mutations) in the allocated space. The execution of individuals is carried out in parallel, i.e. all are executed together, at the same time.

Synthetic life, is it possible?

Frankenstein's monster

In the previous two posts in this series we have seen that the generation of synthetic life in the laboratory is probably a process more difficult than some optimists imagine.

Let’s look at one of the latest experiments in synthetic biology: George Church and Nili Ostrov, Harvard biologists, are trying to build a strain of the bacterium Escherichia coli immune to all existing viruses. How? By changing its genetic code so that viruses do not understand it and cannot use the bacterial cellular machinery to reproduce. Since the genetic code is redundant, it is possible to replace one of the codons encoding the amino acid arginine (AGA) with another that also encodes the same amino acid (CGC), and all the genes of the bacterium would go on generating the same proteins. This would be done with several rare codons. But since viruses would continue to use the substituted codons, the bacterial cell machinery would no longer be able to understand the DNA of the virus. This part of the job is almost finished. When it is done, it would also be necessary to eliminate the transfer RNAs of the missing codons and ensure that they are not remanufactured, so that the cellular machinery can no longer use them.

Note that the work done so far is the manipulation of the data recorded in the DNA. It is equivalent to changing the information contained in the hard disk of a computer so that it stops using a certain instruction of the language of the machine, by replacing it with another equivalent instruction. We are still very far from synthetic biology in the strict sense. Will it be possible to synthesize life in the near future?

Synthetic life, near or far?

In the previous post I detailed some recent advances in the field of synthetic biology, and asserted, without saying why, that I don’t think the goal of creating an artificial living cell is as near as some optimistic researchers believe, such as Craig Venter.

To explain why, I’ll make a comparison between a living cell and one of our most complex artifacts: the computer. A computer consists of the following two main parts:

1. CPU (central processing unit): as its name indicates, it’s the control center and the place where programs are executed. One of its fundamental elements is the machine language, a relatively complex binary code that the circuits of the unit interpret and execute. Every program, in order to run, must be written in machine language.
2. Memory. There are several types: hard disk, which stores the programs and data accessed by the computer, including the operating system, although many of them will never be used; cache memory, faster than the hard disk, which stores those programs and data currently being used, to speed up their process; external memories (such as flash memory), used to transmit data and programs from one computer to another, or to save copies in case of loss of information.

Synthetic life, when?

Craig Venter

First, a clarification. We must distinguish two very different fields of research:

1. Artificial life: this is the part of computer engineering that tries to build programs that emulate the behavior of living beings: either artificial living beings, or colonies of living beings, such as anthill or hives.
2. Synthetic life: this is the part of biology that tries to build artificial living cells from simple chemical substances. So far, this goal has not been achieved.

Shall we be able one day to make life in the laboratory? A few important steps have been taken during the last half century.

• In 1967, Arthur Kornberg (1959 Nobel Prize together with Severo Ochoa) used the enzymes DNA polymerase (discovered by him) and DNA ligase to duplicate the DNA of the fX174 virus, which is made of 5386 nucleotides, and showed that the copy of the virus could infect bacteria, as the original virus. For those who argue that viruses are alive, this was the first generation of artificial life, but the authors of the experiment, who did not share that opinion, insisted before the press that their discovery shouldn’t be presented in that way.
• In 1976, Frederick Sanger (the only winner so far of two Nobel Prizes in chemistry, in 1958 and 1980) managed to sequence the genome of the fX174 virus, i.e. obtain the complete ordered list of its nucleotides. This was the first genome successfully sequenced.
• After his spectacular triumph in the Human Genome project, when a small private company achieved results comparable to the multi-million dollar project sponsored by the United States Government, biologist Craig Venter dedicated his efforts to synthetic biology. The first thing he did was build artificially DNA molecules, starting from the list of their nucleotides, and make those molecules act inside living cells as their natural models do. In 2003, Venter and his team built the first “artificial DNA” by generating the DNA of virus fX174, using DNA synthesizer machines, and starting from the list of the virus nucleotides obtained by Sanger. This was not the first artificially constructed virus, since in 2002 Eckard Wimmer and his team had managed to synthesize RNA from the poliovirus that causes polio, starting from its genome (the list of its nucleotides).

• After this achievement, Venter and his team moved to more complex organisms, about which there is no doubt that they are living beings, and to begin with they chose the group of living cells with the smallest known genomes: mycoplasmas, very small bacteria without a hard membrane, which makes their handling easier. The smallest genome belongs to Mycoplasma genitalium and contains 582,970 nucleotides and 480 genes. This genome was sequenced by Venter and his team in 1995, and in 2007 they managed to synthesize it (with some changes, to facilitate its identification) from the list of its nucleotides.
• The next step, completed in 2007, was transplanting the DNA from a bacterium into another bacterium of a related although different species, to see if it could be expressed there. For this they chose two similar species: Mycoplasma capricolum and Mycoplasma mycoides, which have a larger genome with 1,010,023 and 1,083,241 nucleotides respectively, 91.5% of which are the same, which made it likely that the genome of one species would work in a cell of the other. They extracted the chromosome from M. mycoides, inserted it into M. capricolum cells and allowed the cells to reproduce, hoping that some of the daughter cells would be left only with the transplanted genome, as it happened. The DNA taken from M. mycoides was able to reproduce correctly in a cell of M. capricolum, so this cell had changed species.
• Successive research, which ended in 2010, was aimed at artificially synthesizing DNA from M. mycoides and inserting it into cells of M. capricolum, to see if the change of species could be made, not from the DNA of a living cell, but with an artificially generated copy. This experiment also ended successfully.

Venter himself acknowledges that his experiments have not resulted in the synthesis of living cells. In every case, they have started from pre-existing cells whose DNA has been replaced by another DNA, either from a different cell, or artificially generated. To be able to say that life has been manufactured, it would be necessary to design synthetic DNA and introduce it into an artificial membrane, with artificial contents, getting the artificial cell to reproduce. Until this is achieved, the synthesis of life in the laboratory won’t have happened. Although Venter does not risk predicting a date when this will have been achieved, he doesn’t think the goal is too far away.

I think he’s wrong. Perhaps we shouldn't be too optimistic. Remember the horizon effect. In the next post I’ll explain why.

The same post in Spanish
Thematic Thread on Synthetic and Artificial Life: Previous Next

Manuel Alfonseca

The Fifth Level of Evolution

The theory of evolution is well established by scientific evidence, but is far from explaining everything. Some puzzles remain pending whose resolution does not seem to be immediate:

• The origin of life. We do not know how, when and where it happened. There are many theories, but none has been proven and are very difficult to prove, because the origin of life, rather than a scientific fact, is a historical fact. It is not enough if we were able to reproduce it in the laboratory, it’s necessary to find documentary evidence that this is how it happened, not otherwise. It’s very likely that these tests cannot be found, because the paleontological traces of the origin of life have surely been lost.
• The mystery of the change in level. Throughout the history of life on Earth, living things have gone through several successive levels:

Chance or pseudo-chance?

Gregory Chaitin

In computer programming, certain algorithms (called pseudo-random) generate series of numbers that meet the conditions required by statistics to decide on the randomness of a sequence. These algorithms are used frequently to simulate chance.

However, these algorithms have been designed by someone (the programmer who invented them). In fact, they are not usually random, in the sense that, if they are executed several times in a row, they always give the same results.

We have a similar case with the digits of p. Ten trillion digits of p are currently known, and their number is constantly growing. So far, the digits of p have met all statistical randomization tests. However, it is evident that they cannot be truly random, that they are designed. There are simple algorithms that generate them one after another, in the correct order.

Let us go back to the mental experiment of the previous post in this blog. If intelligent beings were to emerge in an artificial life experiment,

Would these beings be able to distinguish between chance and design as the origin of their own existence?

Chance or design?

Tree of life

In this context, we must distinguish three things:

1. The scientific theory of evolution, which is strongly supported by data from other sciences, such as embryology, comparative anatomy, paleontology, biogeography, or molecular biology (DNA analysis).
2. The claim that evolution is a consequence of pure chance, which is not a scientific theory, but philosophical, although its supporters claim that it is scientific.
3. The assertion that evolution is an example of design, which is not a scientific theory either, but philosophical. The supporters of intelligent design argue that it is scientific.

To solve this dilemma we would have to answer one of the following questions:

•         Is there a way to prove scientifically that evolution is a consequence of chance, rather than design?

•         Is there a way to prove scientifically that evolution is a consequence of design, rather than chance?

What is life?

Saccharomyces cerevisiae

After a century discussing about the origin of life, we are not closer to knowing what did happen. In the mid-twentieth century, when Stanley Lloyd Miller performed the famous experiment where he applied energy to a mixture of methane, hydrogen, ammonia and water, and obtained amino acids, scientists announced the imminent manufacture of artificial life in the laboratory. Such estimates are often too optimistic. In this case they were.

The first question to be solved here is what is meant by being alive. If we consider the problem carefully, we’ll find beings that are clearly alive and others that definitely are not. Plants, animals and ourselves are alive. Stones, distilled water, carbon dioxide, are not. In these cases deciding is no trouble. When Antony van Leeuwenhoek discovered microorganisms (yeast, infusorians, bacteria, spermatozoa and red blood cells) nobody doubted that they are alive. But things are not always so simple.

Artificial life is not here

Saccharomyces cerevisiae (yeast)

Last spring, the media published the news that a scientific team had replaced the smallest chromosome of a yeast cell by a synthetic chromosome, built from the nucleic acid sequence of the replaced chromosome with a few changes, such as the elimination of a section. Once added to the yeast genome, the synthetic chromosome seemed to work correctly.

The headline of the article linked above is meaningful: Scientists Move Closer to Inventing Artificial Life. As it is worded, it seems to imply that we are close to building artificial life. But is this true? Or is this one of those typical overstatements of the media?

In the scientific parlance, artificial life may have two very different meanings: