Forgot your password?
Jianpeng Ma(02-25-05)
EMAIL
Document Actions
(10-26-05) pearson correlation substitutes
(10-26-05) pearson correlation substitutes
(10-26-05) pearson correlation substitutes
Dr. Haiyan Huang went here and gave a talk about the poisson-distribution-derived method[4] and another method to remove dependency from datasets and then do correlation calculation.I came up two ideas when reading her paper.
1. For each row, calculate the slope between adjacent columns, and compare the slope vector of each gene. This is used to overcome the fact that pearson only reconizes the changing trend(direction of slope) and don't care about the magnitude of the slope.
2. Use a polynomial(order 4 or 5) to fit the expression vector and compare the coefficient vector. This is contrast to the fact that pearson is weak to curvilinearity.
Yu Huang 2006-03-25
Gene Regulatory Networks Special Feature (PNAS)