doi:10.1038/nindia.2011.67 Published online 16 May 2011
Periodic table is an arrangement of all known elements organized on the basis of structure and atomic number. Moving across the periodic table, one finds an increase in the number of protons. Moving down in the group, the number of electron shells increase. This pattern is so consistent that predicting the properties of elements based on their relative position in the table, is straightforward.
If we use the same tenets to tie all the molecular data to corresponding higher level behaviours, biology would get a flavour of an engineering discipline.
To achieve this goal, one would need a lot of 'relevant data' at different molecular and interaction levels. The fundamental question is: how would the elements of a bio-periodic table look like?
To answer this question, one needs to set some general parameters. An element of a bio-periodic table should ideally be (a) non-redundant, (b) directly involved in the process, (c) available in finite numbers, and (d) connected to common higher level processes and organisms.
The 'protein fold' qualifies to be the element of a bio-periodic table. Protein folds are compact, secondary structure arrangements of the polypeptide chain that optimize packing of residues and enable interactions. The non-redundant nature of the folds can be seen by the fact that protein molecules can adopt same folds even though the sequences may be very different.
Further, due to stereochemical constraints, the number of ways that a sequence can fold is believed to be finite, in the range of several thousands1, 2. Due to this reason, the Protein Data Bank contains a much smaller number of folds than the number of sequences discovered3.The reason for saturation of protein folds beyond a threshold is unclear at the moment.
Why shouldn't we consider genes and RNA as elements of a bio-periodic table? There are several reasons why these molecules may not be ideal. First, it is the sheer enormity of the diversity that one has to encounter in the nucleic acid space. Second, the molecular interaction information is not easily visible from the DNA sequence alone. Third, RNA molecules ferry and regulate information and are not directly involved in running pathways and networks.
Moving upwards from the fold level, the bio-periodic table would enable users to connect fold-description with the high cell-level description through a series of hierarchical information transfers (picture on right). Protein fold represents a kind of 'bottleneck' where information compresses from the 'nucleic-acid-end' and expands at the 'molecular-interaction-end'. Further, it is interesting to observe a nearly regular periodicity of information compression and expansion from DNA to the network level.
Given that interaction is the fundamental feature of elements of a bio-periodic table, the next question is whether 'regulation of interaction' also should be part of the interaction table. Well, the 'regulation space' is enormously wider than the 'process space' as every organism has a unique way to regulate information processing from molecules to networks. It is highly unlikely that all the regulatory mechanisms from all the organisms can be compressed into a single 'interaction management table'.
In its simplest form, an interaction table specific to an organism would topologically connect protein folds with pathways and a regulatory matrix would indicate how this process is managed in that organism. One also needs to consider that quantitative thresholds, contextual environmental and temporal descriptions would be needed to give a sense of completeness to the multi-dimensional regulatory table.
To realize the goal of periodic table, the good old structural biology seems to be more relevant than ever. From interaction data, one can take off to pathway and network level and describe behavioural outcome of the system.
Though it looks like a distant goal, due to challenges in collecting accurate and enough data, the bio-periodic table would be useful in providing well-defined modules based on biological rules of composition.
Once a reasonably good periodic table is in place, both systems and synthetic biology would enormously benefit. One can build reliable theoretical models of processes in biology. Likewise, to assemble new systems from scratch would be relatively straightforward. Given that interaction is the fundamental feature of a functional object, the table would predict the potential impact of knock-outs, knock-ins by provide a transition corridor between genes and pathways.
While physicists and chemists enjoy a number of well-defined rules, principles and laws, biologists must contend with the rules of Mendelian Inheritance. The bio-periodic table would be a good place to understand biological complexity in terms of variables and constants in biology, build whole cell-scale models and attempt in developing a theory in biology.
Despite its obvious advantages, it is important to point to the fact that unlike periodic table in chemistry, a change from one element to the next may not be uniformly incremental i.e. it may not be "periodic" in the sense of a periodic table in chemistry. However, even the classic periodic table has few unsolved glitches. For example, the position of Hydrogen is not absolutely correct as it resembles both the alkali metals and halogens. In the classic version, cobalt with heavier atomic mass comes before nickel.
Since Mendeleev's table does not capture the entire space of regularity among elements, alternative forms of table e.g., spiral, circular, pyramidal have been suggested.
Thus, it is quite possible that the final version of a bio-perioidic table may not look like table. However, if one creates a periodic table like representation in biology, it would enable a new way of doing biology, both in the analysis and the design space.