Microbes learning correlated cues and anticipating vital changes in the environment.
It's increasingly apparent that at least some metazoans haven't yet acquired these skills
Science and Environmental Health Network
In nature, environmental changes such as variations in temperature, pH, and oxygen level do not typically occur in isolation, but in a temporally coupled and nonrandom manner. An organism that is capable of learning correlated cues can then exploit them in order to anticipate vital changes in the environment -- for example by preparing for nutrient fluctuations or mounting protective responses. In a Research Article in the 6 Jun 2008 Science (published online 8 May), Tagkopoulos et al. reported that even simple microbes can learn to exploit predictable sequences of environmental signals and prepare for future environmental changes ( see the related ScienceNOW story by R. Kwok). The researchers first used computer simulations to show that virtual evolving organisms garnered more energy if they could "learn" that certain signals preceded the arrival of food and launch a preemptive metabolic response. The researchers then looked for evidence of this ability in Escherichia coli bacteria and found that upon exposure to a large temperature change (as might occur after transition from the outside environment into the mammalian body), the bugs shift from aerobic to anaerobic respiration in anticipation of the low oxygen levels in the gastrointestinal tract. When the bacteria were instead exposed to high oxygen levels after an increase in temperature, they "learned" to turn off their low-oxygen responses in fewer than 100 generations. An accompanying Perspective by N.S. Baliga highlighted the study.
From the commentary:
[The Scale of PredictionNitin S. Baliga*
The predictability of cellular responses is the basis for applications as diverse as preventive medicine and the reengineering of microbes for biotechnology. At first glance, the diversity of biological systems suggests that they can adopt a seemingly infinite number of behaviors or states. If this were true, it would severely hinder our ability to predict the responses of biological systems to new environmental changes. Fortunately, this is not the case (1). An individual biological system functions optimally within an environmental space (defined by the ranges in any given parameter) in which it has evolved. Furthermore, within this defined space, changes in individual factors (temperature, pH, O2, etc.) do not occur in isolation but in a temporally coupled and nonrandom manner for physicochemical reasons. On page 1313 of this issue, Tagkopoulos et al. (2) offer theoretical simulations and experimental validation to show that even simple microbes can learn temporal interrelationships among changes in environmental factors, and thus can predict and prepare for future environmental changes, a behavior attributed to metazoans. This work not only demonstrates how biological networks (gene, protein, and metabolic) have assembled during evolution, but also explains why decoding these networks into predictive models is a tractable problem.]