I have been trying to use ns.ev.DecisionTree() to evaluate lists of markers provided by collaborators. These marker lists can range from 1 - 50 genes, and I have noticed that with gene lists that are longer (let's say > 10 genes), ns.ev.DecisionTree(combinations = True) will stall as it will struggle to write all the generated combinations to memory. There may be many ways to address this, but one idea I have is allowing users to place a cap on the length of combinations generated, as that can reduce the number of generated combinations that need to be written to disk.
I have been trying to use ns.ev.DecisionTree() to evaluate lists of markers provided by collaborators. These marker lists can range from 1 - 50 genes, and I have noticed that with gene lists that are longer (let's say > 10 genes), ns.ev.DecisionTree(combinations = True) will stall as it will struggle to write all the generated combinations to memory. There may be many ways to address this, but one idea I have is allowing users to place a cap on the length of combinations generated, as that can reduce the number of generated combinations that need to be written to disk.