ES_DISCOVER_PERTURBATION - Most likely expression perturbation that creates a given state change [u_ratio_pred, s_ratio_pred, pred_effect] = es_discover_perturbation(N,W,ind_ext,c_ratio,v_ratio,es_result,expansion,expansion_order,u_sigma_prior,s_known_sigma_prior,s_unknown_sigma_prior) Compute the most likely perturbation of enzyme expression (u) and external metabolites (s) that would give rise to a given differential profile of balanced concentrations c and fluxes j This is, more or less, an inversion of 'es_simulate_perturbation' N, W, ind_ext: structural information about the network c_ratio,v_ratio vectors containing perturbations (no matrices for several samples allowed!) es_result: output from a previous elasticity sampling (function 'sample_model') expansion: {'logarithmic','non-logarithmic'} -> depending on the expansion type, the output refers to logarithmic or non-logarithmic values; expansion_order: 1 or 2 for first- or second-expansion_order expansion output: s_ratio: external concentration ratio vector u_ratio: expression ratio vector