Accurate Chemical Classifier Mimics Insect Nose
Two Germans have built a better chemical classification system by taking their inspiration from insect olfactory organs. Here’s how they described their achievement in an abstract from PNAS:1
The chemical sense of insects has evolved to encode and classify odorants. Thus, the neural circuits in their olfactory system are likely to implement an efficient method for coding, processing, and classifying chemical information. Here, we describe a computational method to process molecular representations and classify molecules. The three-step approach mimics neurocomputational principles observed in olfactory systems. In the first step, the original stimulus space is sampled by “virtual receptors,” which are chemotopically arranged by a self-organizing map. In the second step, the signals from the virtual receptors are decorrelated via correlation-based lateral inhibition. Finally, in the third step, olfactory scent perception is modeled by a machine learning classifier. We found that signal decorrelation during the second stage significantly increases the accuracy of odorant classification. Moreover, our results suggest that the proposed signal transform is capable of dimensionality reduction and is more robust against overdetermined representations than principal component scores. Our olfaction-inspired method was successfully applied to predicting bioactivities of pharmaceutically active compounds with high accuracy. It represents a way to efficiently connect chemical structure with biological activity space.
Readers may want to sniff our 06/27/2005 article about the work of Gilles Laurent, whose lab described in detail the exquisite coding method of insect olfaction in Caltech’s periodical Engineering and Science. Laurent’s work was cited in the references.
The scientists took advantage of an algorithm in insect olfactory systems that processes inputs through a system of correlation and decorrelation, finding general similarities and discriminating specific differences. The system maps an almost unlimited number of inputs into the brain, such that “All information necessary to assign a perceptual quality to a chemical stimulus is present there.” They said that “the application of this framework is not limited to the olfactory domain but can also be efficiently used for virtual screening of a pharmaceutical compound database”; i.e., in machine learning systems. They did not build such a system yet. This work only provides a pathway for inventors in the electronic nose world: “The effects of processing in the virtual antennal lobe on those extensions as well as their impact on classification power provide a tantalizing prospect for future research.”
A little more awe for what a tiny insect antenna accomplishes can be gleaned from the authors’ statement, “The processing scheme we present here provides a simplified model of neural computation in the olfactory system.” They said, “We tried to keep the simulation overhead as small as possible so that the essence of the processing strategies would stay obvious.”
1. Schmuker and Schneider, “Processing and classification of chemical data inspired by insect olfaction,” Proceedings of the National Academy of Sciences USA, 10.1073/pnas.0705683104, published online before print December 10, 2007.
2. Gilles Laurent, “Olfaction: A Window into the Brain,” Engineering and Science (LXVIII:1/2), [summer] 2005, pp. 43-51 (PDF).
If we can pardon the researchers’ logical blunder in thinking that evolution could be expected to produce an efficient method for anything, this is quite a good example of biomimetics in action. Biomimetics looks for efficient design and tries to imitate it. This is design science, not evolution. These scientists saw an effective algorithm in an insect classification system, and by plagiarizing it, achieved a high degree of accuracy.
Insects, of course, have no noses; their olfactory sense is located on their antennae. An insect antenna has to be able to map a huge number of chemical signals to the brain. The way it does it is clever and mind-boggling. It is well worth revisiting the Laurent article, if you can disregard the foul odor of evolutionary assumptions here and there.