from biodivine_aeon import * from pathlib import Path model = BooleanNetwork.from_sbml(Path('t-lgl-survival.sbml').read_text()) rg = model.graph() # Fix input variables according to [4] model.set_update_function("Stimuli", "true") model.set_update_function("IL15", "true") model.set_update_function("PDGF", "true") model.set_update_function("Stimuli2", "false") model.set_update_function("CD45", "false") model.set_update_function("TAX", "false") graph = SymbolicAsyncGraph(model) attractors = find_attractors(graph) print("Attractor sets:", len(attractors)) A1 = attractors[0] A2 = attractors[1] A3 = attractors[2] print("A1:", A1.vertices().cardinality()) print("A2:", A2.vertices().cardinality()) print("A3:", A3.vertices().cardinality())

apoptosis = graph.fix_variable("Apoptosis", True) proliferation = graph.fix_variable("Proliferation", True) cytoskeleton = graph.fix_variable("Cytoskeleton_signaling", True) tcr = graph.fix_variable("TCR", True) ctla4 = graph.fix_variable("CTLA4", True) p2 = graph.fix_variable("P2", True) print("==== A1 Phenotypes ====") print("A1[Apoptosis] is always True:", A1.minus(apoptosis).is_empty()) print("A1[Proliferation] is always False:", A1.intersect(proliferation).is_empty()) print("A1[Cytoskeleton] is always False:", A1.intersect(cytoskeleton).is_empty()) print("==== A2 Phenotypes ====") print("A2[Apoptosis] is always False:", A2.intersect(apoptosis).is_empty()) print("A2[Proliferation] is always False:", A2.intersect(proliferation).is_empty()) print("A2[Cytoskeleton] is always True:", A2.minus(cytoskeleton).is_empty()) print("A2[TCR] is always True:", A2.minus(tcr).is_empty()) print("A2[TCR] is always False:", A2.intersect(tcr).is_empty()) print("A2[CTLA4] is always True:", A2.minus(ctla4).is_empty()) print("A2[CTLA4] is always False:", A2.intersect(ctla4).is_empty()) print("A2[P2] is always False:", A2.intersect(p2).is_empty()) print("==== A3 Phenotypes ====") print("A3[Apoptosis] is always False:", A3.intersect(apoptosis).is_empty()) print("A3[Proliferation] is always False:", A3.intersect(proliferation).is_empty()) print("A3[Cytoskeleton] is always True:", A3.minus(cytoskeleton).is_empty()) print("A3[CTLA4] is always True:", A3.minus(ctla4).is_empty()) print("A3[CTLA4] is always False:", A3.intersect(ctla4).is_empty()) print("A3[P2] is always False:", A3.minus(p2).is_empty()) complexAttr = A2.union(A3) alwaysTrue = [] alwaysFalse = [] for variable in model.variables(): isTrue = graph.fix_variable(variable, True) isFalse = graph.fix_variable(variable, False) if complexAttr.is_subset(isTrue): alwaysTrue.append(rg.get_variable_name(variable)) if complexAttr.is_subset(isFalse): alwaysFalse.append(rg.get_variable_name(variable)) alwaysTrue = sorted(alwaysTrue, key=str.casefold) alwaysFalse = sorted(alwaysFalse, key=str.casefold) print("Always *true* in complex attractors:", alwaysTrue) print("Always *false* in complex attractors:", alwaysFalse)

A1_weak = reach_bwd(graph, A1) A2_weak = reach_bwd(graph, A2) A3_weak = reach_bwd(graph, A3) all_states_card = graph.unit_colored_vertices().vertices().cardinality() A1_strong = A1_weak.minus(A2_weak).minus(A3_weak) A1_strong_card = A1_strong.vertices().cardinality() print("A1 strong basin:", (A1_strong_card / all_states_card) * 100.0, "%") A1_realistic_strong = A1_strong.minus(apoptosis) A1_realistic_strong_card = A1_realistic_strong.vertices().cardinality() print("A1[Apoptosis = false] strong basin:", (A1_realistic_strong_card / all_states_card) * 100.0, "%") A2_strong = A2_weak.minus(A1_weak).minus(A3_weak) A2_strong_card = A2_strong.vertices().cardinality() print("A2 strong basin:", (A2_strong_card / all_states_card) * 100.0, "%") A3_strong = A3_weak.minus(A1_weak).minus(A2_weak) A3_strong_card = A3_strong.vertices().cardinality() print("A3 strong basin:", (A3_strong_card / all_states_card) * 100.0, "%") A2_A3_strong = A2_weak.union(A3_weak).minus(A1_weak) A2_A3_strong_card = A2_A3_strong.vertices().cardinality() print("A2/A3 strong basin:", (A2_A3_strong_card / all_states_card) * 100.0, "%") all_three = A1_weak.intersect(A2_weak).intersect(A3_weak) print("Undetermined:", (all_three.vertices().cardinality() / all_states_card) * 100.0, "%") all_phenotypes = A1_weak.intersect(A2_weak.union(A3_weak)) print("Undetermined phenotype:", (all_phenotypes.vertices().cardinality() / all_states_card) * 100.0, "%")

# Reload the graph and recompute attractors without fixing parameters. model = BooleanNetwork.from_sbml(Path('t-lgl-survival.sbml').read_text()) rg = model.graph() graph = SymbolicAsyncGraph(model) attractors = find_attractors(graph) print("Attractor sets:", len(attractors))

healthy = graph.empty_colored_vertices() diseased = graph.empty_colored_vertices() apoptosis = graph.fix_variable("Apoptosis", True) for attr in attractors: healthy_colors = attr.intersect(apoptosis).colors() diseased_colors = attr.minus(apoptosis).colors() assert healthy_colors.intersect(diseased_colors).is_empty() healthy = healthy.union(attr.intersect_colors(healthy_colors)) diseased = diseased.union(attr.intersect_colors(diseased_colors)) print("Parametrisations with healthy:", healthy.colors().cardinality()) print("Parametrisations with diseased:", diseased.colors().cardinality())

cd45 = graph.fix_variable("CD45", True) il15 = graph.fix_variable("IL15", True) # This number should be 48, since that is the number of parametrisations # where cd45 or il15 are true. print(diseased.intersect(cd45.union(il15)).colors().cardinality(), "colors have diseased attractors.")

proliferation = graph.fix_variable("Proliferation", True) assert healthy.intersect(proliferation).is_empty() assert diseased.intersect(proliferation).is_empty()

# Reload the graph and recompute attractors without fixing parameters. model = BooleanNetwork.from_aeon(Path('t-lgl-proliferation.aeon').read_text()) rg = model.graph() graph = SymbolicAsyncGraph(model) attractors = find_attractors(graph) print("Attractor sets:", len(attractors)) healthy = graph.empty_colored_vertices() diseased = graph.empty_colored_vertices() apoptosis = graph.fix_variable("Apoptosis", True) proliferation = graph.fix_variable("Proliferation", True) for attr in attractors: attr = attr.intersect(proliferation) healthy_colors = attr.intersect(apoptosis).colors() diseased_colors = attr.minus(apoptosis).colors() assert healthy_colors.intersect(diseased_colors).is_empty() healthy = healthy.union(attr.intersect_colors(healthy_colors)) diseased = diseased.union(attr.intersect_colors(diseased_colors)) print("Parametrisations with healthy attr. and proliferation:", healthy.colors().cardinality()) print("Parametrisations with diseased attr. and proliferation:", diseased.colors().cardinality())