ES_COMPARE_ENSEMBLES - Compare an output function between two model ensembles
res = es_compare_ensembles(output1, output2, n_per, fdr, verbose)
Significance test for model outputs obtained from multiple elasticity sampling
Significant differences are computed from the outputs of a previous multiple
elasticity sampling for two model variants (see 'es_sample_multiple').
Both model variants must have the same number of metabolites and reactions.
Inputs
output1 - Matrix or tensor from 1st model variant (from 'es_sample_multiple')
output2 - Matrix or tensor from 2nd model variant (from 'es_sample_multiple')
n_per - Number of samples in permutation test
fdr - False discovery rate
Output: Data structure 'res' with fields:
res.mean_total - Mean value for all samples (total mean)
res.mean_a - Mean value for 1st variant
res.mean_b - Mean value for 2nd variant
res.mean_delta - Difference between values for both variants
res.p_value_mean_total - p value of [mean value for all samples ~= total mean]
res.p_value_mean_a - p value of [mean value for 1st variant ~= total mean]
res.p_value_mean_b - p value of [mean value for 2nd variant ~= total mean]
res.p_value_mean_delta - p value of [difference between values for both variants
- ~= difference between values from shuffled variants]
res.mean_total_significant - Significant elements for mean value for all samples
res.mean_a_significant - Significant elements for mean value for 1st variant
res.mean_b_significant - Significant elements for mean value for 2nd variant
res.mean_delta_significant - Significant elements for difference between variants
- Values: 1 (significantly high), -1 (significantly low)
res.n_mean_total_significant - # significant elements for mean value for all samples
res.n_mean_a_significant - # significant elements for mean value for 1st variant
res.n_mean_b_significant - # significant elements for mean value for 2nd variant
res.n_mean_delta_significant - # significant elements for difference between variants