Filtering ASVs/OTUs via mutual information-based microbiome network analysis

BMC Bioinformatics

Table 1 Microbiome network characteristics for different hard thresholds \(\tau\)

\(\tau\)	\(R^2\)	\(\bar{k}\)	\(-\gamma\)
0.05	0.71	5.96	− 0.85
0.10	0.86	3.65	− 1.00
0.15	0.75	2.30	− 0.56
0.20	0.75	1.87	− 0.64
0.25	0.84	1.61	− 0.60
0.30	0.84	1.39	− 0.61
0.35	0.70	1.17	− 2.68
0.40	0.74	1.04	− 0.69
0.45	0.97	0.70	− 0.97
0.50	0.56	0.43	− 4.53
0.55	0.56	0.22	− 3.90
0.60	0.57	0.17	− 3.96
0.65	0.56	0.17	− 3.96
0.70	0.56	0.17	− 3.96
0.75	0.56	0.17	− 3.96
0.80	0.56	0.17	− 3.96
0.85	0.56	0.17	− 3.96
0.90	0.56	0.17	− 3.96
0.95	0.57	0.17	− 3.96

The second column contains coefficient of variation \(R^2\) that varies between 0 and 1, where 0 indicates that the power-law model explains none of the variability of the empirical degrees, while 1 indicates the model perfectly fit the data. \(\bar{k}\) and \(-\gamma\) are the average node degree and the exponent of the power-law distribution, respectively. Bold values indicate the chosen level of thresholding (highest R²)

ISSN: 1471-2105