Extended Local Similarity Analysis (eLSA) of Biological Data
Local association analysis; Local similarity analysis
The advances in high-throughput low-cost experimental technologies have made possible time series studies of hundreds or thousands biological factors simultaneously. The availability of such datasets leads to an increased interest in profile similarity analysis techniques that can identify significant association patterns possibly embracing biological insights. In the context of metagenomics, factors of particular interest are operational taxonomic units (OTUs), microbial genomes, and environmental genes. Their association patterns may suggest microbe-environment, symbiotic relationships, and other types of interactions.
Many computational or statistical approaches exist to study the profile similarity at global scale, such as Pearson’s correlation coefficients (PCC), Spearman’s correlation coefficients (SCC), principal component analysis (PCA), multidimensional scaling (MDS), discriminant function analysis...