ISOpureS2.model_core.new_model function

Compute loglikelihood given all model parameters for step 2

Produces a list (the model) which initializes the parameters vv, log_BBtranspose, PPtranspose, kappa, theta, omega, log_all_rates for step 2

ISOpureS2.model_core.new_model(tumordata, kappa, INITIAL_VV, PPtranspose, BBtranspose)

Arguments

• tumordata: a GxD matrix representing gene expression profiles of tumor samples
• kappa: a 1xD matrix which represents strength parameter kappa over cc, given the reference profile mm
• INITIAL_VV: a vector with K components, the prior over mixing proportions, theta, with last entry weighed more heavily
• PPtranspose: the prior on the tumor-specific cancer profiles is just the reference cancer profile (1xG matrix) learned in ISOpureS1, standardized so that all entries sum to 1
• BBtranspose: a (K-1)xG matrix of the standardized normal profiles, so that they sum to 1

Returns

• model: a newly generated model list to hold all the parameters

Author(s)

Gerald Quon, Catalina Anghel, Francis Nguyen