"Monoallelic Receptor Gene Expression and Optimal Sensory Coding in the Olfactory Pathway"
In mammals the olfactory system combines both high sensitivity to an enormous range of odours together with high selectivity to behaviourally important compounds. This performance is underpinned by the largest family of olfactory receptors (approx. 1,000 ORs in mice, taking up 3%†) of the genome - each OR having a unique receptive field (RF) to a wide range of odorants. For each olfactory sensory neuron, locally determined stochastic gene regulation processes ensure that a single allele is expressed from this OR superfamily. How does this local stochastic gene selection process lead to high levels of system performance globally? By applying an information-theoretic measure to RF simulations, we have shown that to accurately detect the very high-dimensional stimuli required of the olfactory system, a population of receptors with mixed RFs is optimal. Moreover, with increasing dimensionality of the stimulus, this mixed-RF solution further outperforms the one-receptor one-ligand solution. Our simulations show that randomly selecting between a fixed population of ORs with a diversity of RFs provides optimal detection performance overall. Local stochastic gene regulation processes are therefore consistent with globally optimal detection performance in the mammalian olfactory pathway.
* work done with Manuel A. Sánchez-Montañés, Universidad Autónoma de Madrid catalogued at senselab.med.yale.edu/OrDB
Ort: INB Seminarraum
