Similarity evaluation on noisy time series gene expression data using particle filter and longest common subsequence

Correct and accurate analysis of the similarity score using noisy time series gene expression data plays important role in understanding gene level biological systems. In this study, we propose a two-step algorithm using particle filter and longest common subsequence to compute the similarity score of noisy time series gene expression data from synthetic model and microarray measurements. Specifically, an iterative algorithm is proposed where particle filter is applied to process non-Gaussian noise data and longest common subsequence(LCS) is applied to calculate the similarity score using processed data from the first step. The proposed algorithm can obtain accurate results with unknown noise statistics. The effectiveness of the proposed algorithm is demonstrated using data from the synthetic model and experiments.