Team:Imperial College/Collaborations

IGEM COLLABORATION

Collaboration played an integral role in our project. IGEM teams are part of our target market for SOAP Lab, so in order for our project to remain informed based on our end-users, we needed a strong line of communication with those users. We therefore established a network with a number of teams, who trialled the various aspects of our product as they were in development. 
Additionally, as part of the Education aspect of our project, we mentored five teams on the modelling, which later culminated in our Introduction to Mathematical Modelling in Synthetic Biology package. 
We are very grateful to all the teams we collaborated with, as they were essential in the progression and development of our project.

 Modelling mentorship

Gabe, one of our mathematical modelers, participated in the IGEM 2017 competition with his high school, Judd_UK. At the time, he had very limited knowledge of mathematical modelling, and there weren’t any resources accessible to high school students to learn from either. Thanks to his mentorship from Oxford University however, he was able to formulate a thorough model which had a significant impact on his team’s project, winning them the award for Best Model.  In this spirit, we wanted to give back to the IGEM community and offer up our team’s knowledge of modelling to assist in the IGEM 2020 competition. We recognize it can be such a powerful tool for teams and even a little bit of guidance can be so useful. Through IGEM meetups and the global slack, we collaborated with a total of five teams, mentoring them throughout the summer on their team’s model  

Gaston Day School

Korea HS

University of Lausanne

University of Hamburg

City of London School 

While speaking to our mentees, we realized a commonality between the hurdles they needed to overcome to develop a mathematical model, and we considered how many other teams are struggling now, as well as future teams! So to make modeling more accessible for all teams, current and future, we decided to make an
Introduction to Mathematical Modelling in Synthetic Biology package.

beta testing

stage 1 - Linkoping

One of the driving features of our web tool is the ease of use that comes with a simple user interface and abstracts away all the technical details behind the automatic liquid handler script generation. Early in the project, we got in contact with another iGEM team that were developing their own software tool and had already come quite a far way along. LiU iGEM Linköpings helped us think about how we can beta test our tool and dig into the value that each feature provides, thinking critically about whether it is helping or hurting. With their tool at around 70% completion and ours at around 10%, the beta testing and bug report forms they sent us helped us ask the right kinds of questions further down the line when we were able to do our own beta testing.
Have a look at the following material: BETA TESTING PRESENTATION
BETA TESTING WEB TOOL

beta testing

stage 2 - Paris_Bettencourt

The iGEM Paris Bettencourt team allowed us to put this into practice. We observed their process of making an SBOL document with their parts and identified key points of confusion thanks to the pointedness of our questions and elements of the Beta testing feedback form that Linköpings created. We identified technical issues with them such as scrolling and loading pages, defining the directionality of parts, and getting and saving parts from other working documents in the same session. The combinatorial design feature was initially unintuitive for them to access, but once they were explained how to find and use the feature, we were happy to see that they could set off to do their own combinatorial designs without further input from us. Because of this, we decided to emphasise any help regarding the combinatorial design feature in the SBOL Designer part of the page, aiding the user friendliness of our software.

beta testing

stage 3: Hamburg

Hamburg, our Partner team, were our third and final beta testers. Their team trialled all aspects of our product MVP in October, from designing their parts all the way to running the scripts on the Opentrons. Not only was their feedback throughout invaluable as a full test case for our product, but proving that another iGEM team could implement our pipeline in an afternoon made an incredible case its the trustworthiness. Visit our Partnership page to find out more.

Gaston day school

Gaston Day School's preliminary model made early in the summer

We reached out to Gaston_Day_School after the Germany meetup when they announced they were looking for guidance on their team’s model. We provided guidance on gene expression modeling and general Synthetic Biology modelling ideas, such as whole-cell models, cell-free modeling and the benefits and limitations of modeling. We remained in contact throughout the summer as they developed their code, where we provided guidance on integrating their various models into one comprehensive model.

They were our first mentees, so were the initial test case for our mentorship programme and later, our Introduction to Modelling package. They gave feedback on the teething issues, primarily determining what sort of audience we should aim for our package to be accessible for High School students, and the needs of that audience. They also indicated that the practical application of models was something they weren’t sure about, so we sought to include a modelling in practise section section.

"The code [Imperial_College] wrote for us definitely provides a basic framework for our gene expression and enzyme kinetics model. We would not know where to start our model without [their] guidance."

 Hide

Korea HS

Screenshot of one of our tutorials. Gabe is talking through some example code he produced for their model

We reached out to Korea_HS through the Global Slack. They were building off of their previous team’s project and were having trouble seeing what they could model, and what purpose a model would serve for their project. As well as introducing the basic gene expression model, we gave bespoke advice as to how the results of the model could influence their project, specifically by suggesting they should find an ideal region for the concentration of their expressed protein within which it functions as they require it to.We met with this team a number of times, providing guidance throughout their project, from gene expression, to binding, to modelling in practise. We also feel their continuous input to our Introduction to Modelling package was invaluable in its development. They read through our documents and watched our tutorial videos in detail, providing feedback on the finer details of both. We therefore feel this collaboration satisfies the partnership requirements as well.

Screenshot of one of our tutorials. Gabe is talking through the graphs this code makes

"The Imperial College team provided excellent advice and detailed suggestions on how we can proceed with our project's mathematical modeling. Our team reviewed their mathematical modeling package and it was like no other!"

 Hide

Unilausanne

Graph produced from the example code given to model a repressilator system

We met with UNILausanne at the German meetup, and later reached out to them, as they were having trouble with their modeling. Their project is based on the repressilator system. We introduced a basic gene expression model for a general repressilator system and the code needed to simulate the model. They wanted to slow down the oscillation time period of the repressilator to 24 hours using a binding agent. We suggested using the hill equation or integrating the binding as an ODE into the simulation in order to test the effectiveness of such an addition, and find the conditions needed to extend the period to 24 hours. We also provided 
CODE to get them started, which they built upon by using their own constants, and later, they integrated a stochastic population growth model alongside the gene expression model. 

After our first meeting, it became clear that the hill equation is not an intuitive equation to use. We therefore felt a derivation would be necessary for a deeper understanding of the dynamics and assumptions made. The team really appreciated the derivation and so we determined it was a necessity to include it in our Introduction to Modelling package. We also determined that stochastic modelling, although much more complex than deterministic modelling, would make a great addition to the package for teams with a slightly stronger foundation in modelling, primarily undergraduate teams. The starting code also enabled them to simulate their models, so we determined that coding should be incorporated as well.

"We are really thankful to ICL iGEM team for introducing to us the basic concepts of modelling. Given our little experience in modelling, our dry lab team needed some advice in different aspects; in particular, their video calls, written tutorials, and python scripts helped us to better understand ODE models as well as the Hill function and its derivation from Mass Action kinetics. They were always nice and supportive!" 

 Hide

University of Hamburg

We began our partnership with Hamburg with them participating in our mentorship scheme. They provided feedback on our Introduction to Modelling package, but were keen to help us out further. Given we were having a lot of trouble getting in the lab, and they had access to Opentrons, they were ideal candidates for a prototype test case for our software tool. Find out more here PARTNERSHIP 

In our first meeting, they presented their project and the dynamics of their system. We then gave a tutorial on the basics of modelling in Synthetic Biology. We first described their system using a flow diagram, illustrating transcription and translation, the interactions between different molecules and degradation. We then elaborated on how they could use Mass Action kinetics, or more generally a Deterministic modelling framework, to model this system. We suggested assuming the ribozyme cleaving process was a simple, instantaneous reaction to begin with, enabling further complexity to be added. Finally, We followed up with some code which could simulate their system, using arbitrary constants. Below is an example of the graphs which this 
CODE can produce:

Graph produced from the example code to simulate their system

In a follow up meeting, we discussed compartmentalisation, as in their system, they have two compartments with a partially permeable membrane between them. We suggested using a simplified version of Fick's law to do this, discussing the assumptions and limitations and how complexity could be added.Hamburg was an excellent test case for our Introduction to Modelling. They pointed out the strengths and weaknesses, which we carefully considered and integrated into the package. Their inclusion of compartmentalisation drove us to include it in our Introduction to Modelling package, since it is one of the key additions which develops a model closer to a whole-cell model. 

"As entering a wet lab was quite difficult this year, the modelling of our ideas was very important. But as none of our team members has modelling experience, we had some difficulties modelling our project. Fortunately, we had the opportunity to start a collaboration with the iGEM team from Imperial College London - at a perfect time point when we were all desperated about designing our model. We met several times via online conferencing platforms with the team and realized that their idea of developing a modelling starter pack suits perfectly with our problem. This is how we got great help in writing the code to describe our project in mathematical terms. Moreover, they wrote a manual describing most of the equations needed in the model and helped us to understand the whole process. They also helped to alter the code for the different project scenarios."

 Hide