The repository contains tools for data generation and analysis for the eLife reserach article - "The solubility product extends the buffering concept to heterotypic biomolecular condensates" (https://elifesciences.org/articles/67176)

1. NFsim simulations

• We generate the two components bngl files (http://bionetgen.org/) using a custom python script (file 00); three components files (A₃ - B_1,3 - C₆ system) are written manually.
• The bngl file (file 01) is then converted to an xml file (file 02); detailed methods can be found here https://emonet.biology.yale.edu/sites/default/files/files/NFsim_manual_v1_11.pdf.
• The xml file is then used as an input to run multiple stochastic trajectories. The jobs are parallely executed in a high performance computing facility (https://health.uconn.edu/high-performance-computing/) using a custom shell script (file 03).
• Each trajectory or run would generate two output files:
  • gdat file containing the observables data (Run_x.gdat)
  • species file containing the molecular clusters at the last time point of simulations (Run_x.species)
• Once we have multiple stochastic runs, we use another python script (file 04) to compute two variables:
  1. average counts of obvervables (Free or Total molecular concentration) across multiple runs
  2. distributions of molecular clusters at the last time point (note that the distribution does not correspond to a steady state when total concentration diverges with time in "CMC" simulations)
• For "FTC" method, we gradually increase molecular counts (100, 200, 300, ..., 1000) at fixed Kd (3500 molecules) to titrate up the concentrations; creation and decay rates are set to zero to make it a closed system.
• On the other hand, for "CMC" method, we begin with 1 molecules each (Kd = 3500 molecules) and increase the creation rates (1290000, 1300000, 1310000 molecules/s) with a fixed decay rate (10000 1/s) to titrate up the clamped concentrations. File_05 is an example for CMC model for 4v-4v system.
• MixedValent system: File_06 and File_07 are examples for 5v-3v and A₃ - B_1,3 - C₆ systems respectively.

2. SpringSaLaD simulations

• SpringSaLaD models are built using the graphical user interface (GUI). The jar file and detailed tutorail can be found here https://vcell.org/ssalad.
• After defining the molecules and reaction rules, volume of the simulation box and simulation time (along with timestep and output inteval time) are set. File_08 is an example input file (for the "reference system") containing all the simulation details.
• SpringSaLaD has funtionalities to launch multiple simulations locally from the GUI. But due to the intensive computations, we run multiple simulations (file_09) in our hpc facility. File_10 is the numerical solver needed to execute such simulations.
• Once multiple runs are executed, a python script is used to compile all those data and perform statistical analysis of the molecular concentrations (file_11) and molecular clusters (file_12).