def generateRandomProblem(nbGenes, nbSpecies, orthologProb = 0.5, paralogProb = 0.4):
""" Generates a random set of genes, orthologs, paralogs and species tree, ready to be input in ConstraintGraph class.
Returned value has the form
{ "genes" : geneset, "orthologs" : orthologs, "paralogs" : paralogs, "speciesTree" : speciesTree }
geneset items have the form [GENENAME]:[SPECIESNAME]
Gene species are attributed randomly, though each species has at least one gene.
If two genes have the same species, they end up in paralogs, always.
:argument nbGenes: Number of genes to generate
:argument orthologProb: chances for 2 genes to be a pair in orthologs
:argument paralogProb: chances for 2 genes to be a pair in paralogs
"""
speciesnames = range(0, nbSpecies)
speciesTree = TreeNode()
speciesTree.populate(nbSpecies, speciesnames)
for node in speciesTree:
if not node.name is None:
node.name = str(node.name)
genes = []
#first add one gene per species
for i in range(nbSpecies):
genes.append("g" + str(len(genes)) + ":" + str(i))
#then fill in the rest with random species genes
paralogs = set()
orthologs = set()
for i in range(nbGenes - nbSpecies):
s = random.randint(0, nbSpecies - 1)
genes.append("g" + str(len(genes)) + ":" + str(s))
#and here we decide of random relationships
paralogProb += orthologProb
for i in range(nbGenes):
g1 = genes[i]
pz = g1.split(":")
g1name = pz[0]
g1species = pz[1]
for j in range(i + 1, nbGenes):
g2 = genes[j]
px = g2.split(":")
g2name = px[0]
g2species = px[1]
if g1species == g2species:
paralogs.add( (g1, g2) )
else:
p = random.random()
if p < orthologProb:
orthologs.add( (g1, g2) )
elif p >= orthologProb and p < paralogProb:
paralogs.add( (g1, g2) )
geneset = set()
geneset.update(genes)
return { "genes" : geneset, "orthologs" : orthologs, "paralogs" : paralogs, "speciesTree" : speciesTree }
