How do you handle a situation where experimental data conflicts with your hypothesis at 10x Genomics?