Neurodegenerative disorders are a set of slowly evolving neuropathological conditions with progressive loss of function and eventually death of neurons participating in motor, sensory or cognitive functions, including for example Alzheimer's, Parkinson's, Huntington's diseases and amyotrophic lateral sclerosis. Brain connectome biomarkers are playing an increasingly crucial role in identification and prediction of neurodegenerative disorders (https://doi.org/10.1016/j.neuron.2012.03.004; https://doi.org/10.1016/j.neuron.2019.08.037). However, the molecular underpinning of these identifications and predictions is still unclear to date. The recent construction of brain transcriptomic atlases has made it possible to explore the relationship between these brain connectome biomarkers and gene transcription (https://doi.org/10.1016/j.tics.2018.10.005). This project aims to identify the potential transcriptomic pattern underlying these brain connectome biomarkers using machine learning models. Elucidating these transcriptomic signatures will benefit our understanding of the fundamental principles of neurodegenerative pathogenesis, and will provide a promising road map for future neurodegenerative research.