A Massachusetts General Hospital-led team has developed a machine learning model that can predict whether mutations affecting ion channels will cause disease. Variations in genes encoding proteins involved in voltage-gated sodium and calcium channels are associated with a variety of mostly neurological and nondevelopmental diseases ranging from autism to heart conditions, but experimental studies of these variants are laborious. To address this issue, the researchers developed a model that they trained on sequence- and structure-based protein features to characterize genetic variants in sodium and calcium channels as loss-of-function versus gain-of-function or neutral versus pathogenic. As described in Science Translational Medicine, the model was applied to exome-wide data from 21,703 cases and 128,957 controls, making predictions that corresponded to molecular loss-of-function versus gain-of-function effects for 87 functionally tested variants.
Using a combination of forward genetics and in vivo electrophysiology, investigators from the University of Oxford and University College London have identified a mutation in mice that disrupts the animals' ability to sleep normally. As reported in Science Advances, the scientists used video tracking to measure sleep in mice, identifying a pedigree characterized by a substantially reduced propensity to transition between wake and sleep states, especially in terms of initiating rapid eye movement sleep episodes. Using whole-genome sequencing, they identified a mutation in the coding sequence for the synaptic vesicular protein VAMP2, which is involved in synaptic vesicle fusion and neurotransmitter release. Further investigation using electroencephalographic and electromyographic methods revealed that the affected mice have a diminished capacity to switch states of vigilance, possibly resulting from an altered excitability balance within local circuits controlling sleep-wake architecture.