Sample size was well known to influence edge weight in correlation network. We plan to construct networks for A (n=152) and B (n=30) groups of samples (human gut microbiome). Do you think we should correct the correlation coefficient with their different sample size?