Vintage biological frameworks: Robert Rosen (Later, Kauffman, Waddington)

It’s fun to be able to know the future — the trick is to be in the future. For example, we can go back to text on how biology was considered almost fifty years ago, and see how it’s developed since.

In Rosen’s “Cells and senescence” (1978), he asks three questions:

1. Do individual cells undergo senescence?
2. If so, does cellular senescent suffice to explain senescence of the organism?
3. If not, what cellular properties are involved in the senescence of the organism?

For 1) “All that’s been experimentally established is that cells can exhibit clonal aging”.

• Sounds still true

So 2) becomes does clonal aging suffice to explain senescence of the organism?

So, how have things progressed since? A near half-century later, many claims seem untouched:

“It is easy to see that the overwhelming bulk of the biochemical and ultrastructural investigations reported above have to do almost entirely with the states of individual cells.”

• single cell omics

“And as we have seen, network aspects suggest very strongly that it is possible for a system of interacting elements to fail in some global sense without any component subsystem failing. Stated another way, we can say that senescence studies concentrating entirely on cell states view the phenomenon entirely in terms of local subsystem failures; taking network considerations into account suggests that the study of senescence cannot be simply a study of such local failures.”

• Single cell omics, senolytics, cell reprogramming

“Another pertinent remark is that, in networks, it is difficult to disentangle cause from effect. Indeed, in any system with a feedback loop, in which a sample of the present behavior is fed back into the system, an initial cause itself becomes an effect, and conversely. It is clear that any mode of analysis which ignores the loop will be hard put to understand the observed behavior of the system as a whole. This is perhaps the ultimate reason why we observe the phenomenon mentioned above, that every experimental observation can be made compatible with every theory; in a system with many loops, any pair of processes may eventually become coupled, so that either can be regarded as a cause of the other.”

• Correlational/regression analyses, unresolved possibility space of theory

“Clearly, there are no fewer variations of programmed senescence theories than there are of error theories. There are also numerous combinations and intermediate theories involving both error and programming, which may be plausibly entertained.

The situation is not improved by comparing any of the theoretical models described above with the results of experiment and observation. There is sufficient flexibility, on both the experimental and theoretical sides, to conclude that at present every experimental observation can be made compatible with every theory…It is not clear that this situation will be improved by further experimental work alone; it is more likely that quite the opposite will occur. This is a most unsatisfactory state of affairs; we have to consider whether the experimental literature truly reflects a totally heterogeneous biological situation about which no generalizations can be made, or whether something essential is missing from past considerations of senescence.”

• Continued debate between these theories; seems similarly unresolved

So, how to investigate these non-component/network aspects? “We believe that such systems cannot be understood without extensive use of carefully chosen model systems, perhaps initially even of a nonbiological character. For instance, a study of the establishment and failure of coherent behavior in populations of coupled oscillators should be most instructive in determining the respective roles of local lesions (i.e., gross changes in state in individual elements) and global network modifications in the loss of population characteristics. The growth of intuitions concerning network aspects in populations arising from the study of such model systems may help us understand what to look for in seeking to understand senescence of organisms in terms of the cell theory, or understanding clonal aging in terms of biochemistry.”

To consider/mull over further:

“As is well known (e.g., Shannon and Weaver, 1949; von Neumann, 1956), redundancy is conceptually the only way to defeat error in communication and in algorithmic systems. In such literature, much attention is given to the artful application of redundancy to increase reliability, that is, to make the error-free life span of a composite system greater than those of the fallible components which comprise it. It is one of the ironies of senescence research that cellular theories of senescence often seek to make the composite system have a shorter life span than its constituent parts. The trivially simple considerations outlined above are the one way of expressing the war between proliferation (the creation of redundancy) and failure.”

“It should also be noted that Franks et al. (1970) were unable to maintain normal human epithelium in culture when it was separated from its stroma but could maintain undissociated tissue in culture for years; and Gey et al. (1974) report that chick fibroblasts could be maintained for years on a collagen substrate. The independence of the Hayflick limit from conditions of culture must be considered in the light of such results.“ https://onlinelibrary.wiley.com/doi/10.1002/path.1711000206 https://pubmed.ncbi.nlm.nih.gov/4817732/

Xxx What makes a complex system nonsimulatable? Too large of a computable problem? Asymptotic approximation of chaotic behavior? Anticipatory systems Or, what makes something computable/formalizable?