2014.12.19 Characterizing complex diseases by big biological data in the forms of dynamics and network
题目： Characterizing complex diseases by big biological data in the forms of dynamics and network
Key Laboratory of Systems Biology, Chinese Academy of Sciences
We described a few new network-based methodologies for solving bio-medical problems in a dynamic manner based on big biological data. (1) we developed a new concept, edge-biomarkers, which transforms‘node expression’ data into the ‘edge expression’ data and thus can classify the phenotype of each single sample in the form of network; (2) we proposed a path-consistent analysis method based on the measurements to reconstruct gene regulatory networks, which theoretically can infer the network structure without the approximation even with a small number of samples, which cannot be achieved by the traditional approaches; (3) we derive theoretical results based on a dynamical network biomarker (DNB) that serves as a general early-warning signal indicating an imminent sudden deterioration before the critical transition of a disease occurs; (4) When a system is constantly perturbed by big noise, it becomes a difficult task to identify the early-warning signals of a critical transition due to the strong fluctuations of the observed data. In this work, we present a new model-free computational method based on the observed time-series data even with big noise to detect such warning signals just before the critical transition. The key idea behind this method is a new strategy: “making big noise smaller”, which increases the dimensionality of the observed data and thus makes the noise smaller in the transformed higher-dimension data.
We adopted omics data of several diseases to demonstrate the effectiveness of our works.