Abstract: Some genetic variants have been discovered to associated with cerebrospinal fluid (CSF) in Alzheimer's disease, yet it does not adapt with the small effect risk variants and ignores the inter-relationship among loci and outer-relationship with environment. As an alternative strategy, functional network-based and pathway-based information could be involved in high-level variants identify process, from system biology view. On the basis of genome-wide association analysis, a weighted-adjusted PageRank network function module mining strategy was proposed to analysis CSF t-tau, which was arranged as a measure reflecting the early pathological characteristics of Alzheimer's disease. The strategy can not only excavate widely studied t-tau-related genetic variations, but also can excavate subnetworks enriched in pathways, such as neurodegenerative diseases, nervous system and signal transduction. It indicates that, subnetworks identified by the strategy have a certain functional relationship with t-tau at the system biology level.