Software

Simulating Systems, Guiding Design

Exploring the computational and mathematical models aiding our project. Serving as the blueprint for the development of our project. Helping us validate design assumptions, bridge theory and practice, and gain insights into territories where the flame of experimentation cannot shed its light.

page-centrepiece

BBrickit

This was a legacy software of ICT Mumbai that was made to help biobrick any DNA sequence. We made a small edit to allow the user to check if their DNA sequence already exists as a biobrick they can order by using the IGEM registry, and made improvements to the GUI to make it user-friendly. The backend was improved to derive biobrick from a repository for enhancing accessibility.

Usage:

`python BBrickIt_V3.py`

Make sure to have the XML file in the same directory when downloaded

nitrogen-pathway

REYNA Model

This Python-based computational model was developed following the framework proposed by Belov et al., which describes the induction and repair of DNA double-strand breaks (DSBs) through four principal cellular pathways: non-homologous end joining (NHEJ), homologous recombination (HR), single-strand annealing (SSA), and alternative NHEJ (Alt-NHEJ). The model represents these mechanisms as a series of rate equations solved as an initial value problem, effectively treating the cell as a batch reactor.

The number of unrepaired DSBs remaining after a specified time interval is then used as an input variable in an exponential survival function to predict cellular survival. The survival curve was fitted to real data of normal yeast cells irradiated

nitrogen-pathway

To simulate the radioprotective effect of the Damage Suppressor Protein (DSUP), a protection ratio r was introduced. This ratio scales down the effective DSB yield to represent the observed increase in radioresistance associated with DSUP expression. By substituting the reduced DSB values into the survival model, an enhanced survival percentage is obtained

The protection ratio r was estimated based on experimental findings from studies on DSUP-expressing tardigrades and mammalian cells (Hashimoto et al., 2016; Kirke et al., 2020; Yoshida et al., 2017).

nitrogen-pathway

Fit obtained when the model is run on a timespan of 20 seconds.

nitrogen-pathway

Fit obtained when the model is run on a timespan of 40 seconds.

The 300 and 450 second timespan models take more than a day to get fully executed in real time Parallelisation was considered believing it could significantly reduce the time constraint by using GPU resources.

Usage:

`python run.py`

`popt, pcov, dose, LET, surv = fit_dsup_to_csv(csv_path='dsup_survival.csv') plot_dsup_fit(popt, dose, LET, surv)`

`popt_base, pcov_base, d2, l2, s2 = fit_to_csv(csv_path='survival.csv') plot_reg_fit(popt_base, d2, l2, s2)`

Model Enhancements for DSUP

dsup_shield_factor: which is derived from

In our project, we adapted and expanded the AMMPER framework to simulate DSUP-mediated radioprotection. The DSUP module modifies the nucleus hit probability and DNA damage propagation by dynamically adjusting the effective survival fraction using functions derived from our Reyna model (model_survival_with_dsup, survival_LQ_with_protection_param, etc.). This creates a tunable interaction between protection parameter (r), radiation dose, and LET, allowing us to estimate how DSUP expression changes survival curves and colony kinetics.

Due to DSUP integration the availUnits, cellRepl and brownianMove have been improved for accounting for the same by shielding a fraction of DNA.

The DSUP is hypothesised to work by shielding the strands of DNA and reducing the number of effective DSBs.

AMMPER

The purpose of our topic was to make a biological reactor radiation-resistant. Specifically, we use Saccharomyces cerevisiae (yeast) as a model chassis and introduce two strategies:

  • Dsup expression – shielding DNA from direct and indirect radiation damage, Expressing DSUP (from Ramazzottius varieornatus) in S. cerevisiae.
  • We utilize AMMPER (Agent-based Model for Microbial Populations Exposed to Radiation) to simulate and compare wild-type yeast, RAD51-overexpressing yeast, and DSUP-expressing yeast. Ammper uses experimental data from nasa’s biosentinel programm which is different from our very own reyna model which is based on the 4 DSBs repair pathways proposed by Belov et al.

What is AMMPER?

AMMPER is an agent-based computational framework designed to model microbial populations under radiation exposure. Key features:

  1. Agent-based: Each cell is an object with position, DNA damage state, and repair machinery.
  2. Direct damage simulation: ions/electrons depositing energy in cell voxels.
  3. Indirect damage simulation: water radiolysis generating ROS species.
  4. Cellular processes: movement, replication, damage sensing, repair (homologous recombination, base-excision), and apoptosis.
  5. Population dynamics: Tracks growth, movement, replication, and repair generation by generation.

Usage:

`python AMMPER.py`

Limitations

  1. Simplified biochemistry: uses probabilistic rules and averaged parameters rather than detailed molecular kinetics; uses blackbox models instead of considering the cellular kinetics of DSBs and repair, respectively.
  2. Neglects microgravity effects: fluid mixing and mass transfer modifications are not currently modeled.
  3. Single-species focus: microbial consortia interactions are not included.
  4. Limited parameters: lacks nutrient modelling, biochemical kinetics, and adaptive stress responses – logistic modelling.
  5. Not physics-based: uses simplified geometry for the system and biological cell geometry as well, which is very primitive.
  6. Single-beam assumption: uniform radiation field, no focused microbeam or mixed fields.