Thursday, January 1, 2009

Computers Making Computers?

An interesting article authored by Antoine Danchin from the Pasteur Institut was recently published and is sure to bring forth much discussion.
Bacteria as computers making computers


Various efforts to integrate biological knowledge into networks of interactions have produced a lively microbial systems biology. Putting molecular biology and computer sciences in perspective, we review another trend in systems biology, in which recursivity and information replace the usual concepts of differential equations, feedback and feedforward loops and the like. Noting that the processes of gene expression separate the genome from the cell machinery, we analyse the role of the separation between machine and program in computers. However, computers do not make computers. For cells to make cells requires a specific organization of the genetic program, which we investigate using available knowledge. Microbial genomes are organized into a paleome (the name emphasizes the role of the corresponding functions from the time of the origin of life), comprising a constructor and a replicator, and a cenome (emphasizing community-relevant genes), made up of genes that permit life in a particular context. The cell duplication process supposes rejuvenation of the machine and replication of the program. The paleome also possesses genes that enable information to accumulate in a ratchet-like process down the generations. The systems biology must include the dynamics of information creation in its future developments.

The quantum teleportation experiments have demonstrated that information can be viewed as a fundamental irreducible property of physics (informationalism). Systems biology is moving in that same direction, as viewing cells as computers with machinery and software makes it possible to view information as a fundamental category of nature and all future developments of systems biology can include this concept when looking at cells.

There are many interesting passages in this article. A few of these are going to be highlighted for discussion.

Historically, systems biology follows on from molecular biology, a science based on many concepts more closely linked to arithmetic and computation than to classical physics or chemistry. Molecular biology relies heavily on concepts such as ‘control’, ‘coding’ or ‘information’, which are at the heart of arithmetic and computation. To accept the cell as a computer conjecture first requires an exploration of the concept of information, in relation to the concept of genetic program.

Cellular processes are exquisitely controlled and carried out by remarkable biomolecular machines. The software needed to coordinate these processes is located in a fairly optimal genetic code that is optimized for evolution and maintains its own functional integrity.

The Austrian mathematician Kurt Godel showed that arithmetic (the science of whole numbers) can make statements about itself. To substantiate this remarkable claim, which implies that just manipulating whole numbers with the rules of arithmetic can generate novel information, G¨odel used a simple trick. He coded the words used in Number Theory as integers (e.g. four, which is quatre in French, vier in German and tessera in Greek, can be coded by 4) and used the corresponding code to translate propositions of arithmetic. This generated a large whole number, which could be manipulated by the rules of arithmetic, and after a sequence of operations, this manipulation generated another whole number. The latter could be decoded using the initial code. Godel’s trick was to drive the sequence of operations modifying the initial statement, to lead to a very particular conclusion. When decoded, the manipulated sequence translated into a particular proposition, which, briefly, stated: ‘I am impossible to prove’. In other words, arithmetic is incomplete, i.e. some propositions of arithmetic can be understood as valid; yet they cannot be proven within the frame of arithmetic. But this ‘incompleteness’ can also be seen as a positive feature; it is what allows the creation of new information – in Godel’s case, the statement of a fact of which the world was previously unaware. In his book, Hofstadter showed that the genetic code, which enables the world of nucleic acids to be translated into the world of proteins, which in turn manipulate nucleic acids, behaves exactly as Godel’s code does. This implies that manipulating strings of symbols, via a process that uses a code, can generate novel information. Of course, in the case of nucleic acids and proteins, there is no Godel to drive the process, and no need for one: while Godel knew what he was aiming at, living systems will accumulate information through recursivity, without any design being required. We only perceive a design because the end result is familiar to us, and thus seems more ‘right’ than any other possible result. But what we commonly term the ‘genetic program’ because it unfolds through time in a consistent manner is not a programme with an aim – it is merely there, and functions because it cannot do otherwise.

Why can't the function of the program be to actively manipulate information as a means to an end... self-replication and preservation. Later in the article something similar to this is actually suggested:

The reluctance of investigators to regard information as an authentic category of Nature suggests that, at this point in the present review of the literature, it may still be difficult for the reader to accept that a cell could behave as a computer. Indeed, what would the role of computation be in the process of evolution? We have already provided some elements of the answer to the question: Turing showed that the consequence of the process of computation along the lines he outlined is that his machine would be able to perform any conceivable operation of logic or computation by reading and writing on a data/program tape. Stated otherwise, and in a way that is easier to relate to biology, the machine manipulates information and, because arithmetic is incomplete [as illustrated in the introduction above (Hofstadter, 1979)], it is able to create information. The machine is therefore in essence unpredictable (Turing, 1936–1937), but not in a random way – quite the contrary, in a very interesting way, as lack of prediction is not due to lack of determinism, but due to a creative action that results in novel information. If the image is correct, then it shows that living organisms are those material systems that are able to manipulate information so as to produce unexpected solutions that enable them to survive in an unpredictable future (Danchin, 2003, 2008a).

There we go, organisms can be viewed as entities that are able to manipulate information as a means to an end. Why would it be difficult to accept that cells to behave like computers? Yet, cells are capable of more than computers, e.g. self-replication and autonomous manipulation of information.

A form of endogenous adaptive mutagenesis (EAM) is also being alluded to in the article:
Living organisms are, therefore, infinitely far removed from the clockwork mechanicism that superficial opponents of molecular biology associate with the widespread analytical stance they call ‘reductionism’ (Lewontin, 1993). It is important to emphasize here that, in the Turing machine, the machine is not only allowed to read the program but also to write on it. If, then, the conjecture of the cell as a Turing machine is valid, apparent paradoxes such as the controversial ‘adaptive mutations’ that enable the cell to invent novel metabolic pathways should not be unexpected (Cairns et al., 1988; Danchin, 1988b).

There is also room for drawing parallels between evolution, memetic algorithms and designed molecular docking programs.

Finally, we must note that the algorithmic approach, presented when considering the genetic program as an authentic program in a Turing machine (Danchin, 2003), identifies two completely different levels: the level of the program and the level of the machine.

The article continues to discuss at length the parallels between our own created information processing systems (computers) and molecular processes fundamental to life. The article is sure to provide information for many more interesting blog discussions.

No comments: