synaptic genes

synaptic genes predicts genome-wide candidate synaptic genes in Drosophila melanogaster using an ensemble machine learning approach trained on temporal transcription profiles to prioritize protein-coding genes by probability of synaptic function.

Key Features:

• Ensemble Machine Learning Model: Trains an ensemble of machine learning classifiers on temporal transcription profiles from known synaptic genes.
• Probability Scoring: Assigns a probability score to each protein-coding gene in Drosophila melanogaster indicating predicted involvement in synaptic functions.
• Catalogue of Putative Synaptic Genes: Produces a catalogue comprising 893 genes postulated to be enriched for undocumented synaptic functions.
• Integration of New Experimental Data: Incorporates 79 experimentally identified synaptic genes into the training set to refine predictions.
• Training Scheme Adjustment: Adjusts the training scheme iteratively based on new empirical data to improve predictive performance.
• Retrospective Enrichment Analysis: Performs retrospective analysis showing significant enrichment of newly identified synaptic genes within the original catalogue.

Scientific Applications:

• Candidate Prioritization for Experimental Validation: Narrows the list of candidate synaptic genes to reduce experimental workload and focus validation efforts.
• Neurobiology Research: Enables discovery and study of genes involved in neuronal synapse assembly and function, informing studies of neuronal function and disease mechanisms.

Methodology:

Train an ensemble of machine learning classifiers on temporal transcription profiles of known synaptic genes; assign probability scores to each protein-coding gene in Drosophila melanogaster; incorporate 79 newly identified synaptic genes into the training set and adjust the training scheme; perform retrospective enrichment analysis of predictions.