Running a simulation or a sensitivity analysis

1. Ssh into sherlock. Clone this repository using git clone. Run ml python/3.6.1.
2. Run a parallelized simulation using the following command: python parallel_simulation.py n_runs simulation_type. n_runs indicates how many jobs you want to submit or how many times you want to run the simulation. simulation_type indicates what to run, whether the baseline simulation, indicated with simulation passed in (or a variation with different motif affinity such as simulation_mutated or simulation_strong), or a sensitivity analysis, which can be one of the following: n_flanks, core_affinity, koff_slope, koff_intercept, n_TF, switching_ratio, diffusion, DNA_concentration, local_volume. For each analysis you do, make sure to create a folder titled with the simulation_type you are calling as well as a folder titled simulation_output within it. In addition, create a folder titled txt_output which is where the outputs from all the jobs will be saved so that you can track jobs or see any errors. There are also a few simulation types that are intended to run locally and do not require parallelization. These are mfpt_simulation, which runs a simulation of just MFPT, one_simulation, which runs one simulation with repeat and random values and plots example single molecule traces, and steady_state, which attempts to calculate the steady state of your system.
3. Check on the status, whether the jobs finished and if there were any errors by calling cat txt_output/*. You can also use squeue -u suid to check on whether or not jobs are running.
4. Process a simulation using the following command: python3 process.py n_runs simulation_type. Alternatively, create a directory called download and then you can process all simulations and copy the output plot to that directory using the command: python3 process.sh.