These weeks started with defining and dividing the subprojects. After joining the safety tours on the labs of all chair groups, we were ready to do the first lab experiments. In addition to a transformation of the iGEM Ptet promoter, we tried to add an extra part to make Ptet*. In these weeks, the constructs of the Gate RNA plasmids and trigger RNA plasmids were designed in silico. We also made competent cells for the paper-based constructs.

Besides the experiments in the lab, we also started with our dry-lab projects. When access to the virtual machine was obtained, we started with analysing and testing SwitchMi Designer and started experimenting with the NUPACK software. We also researched possible optimisation algorithms and feed forward loops, after which we made an object function. For our circuit boolean model, we created and simulated AND circuits and created a test to check the production of GFP using this boolean model.

In this week, we cloned the LacZ construct under the T7 and p70a promoters and verified its production. We started with designing the primers to make the TMSD parts and designing the first DNAzymes. A literature study on RNA sponges was started as well. The AND and OR gate plasmids were made with overlap extension PCR and golden gate cloning. This method was also used to make our first toehold switch (CLSB-UK), and the same was tried for the trigger plasmids for the toehold gates. In addition, we made heat shock competent "Marionette" E.coli cells, which later were used for in vivo experiments. This week we were also ready to sequence the modified Ptet promoter. When we saw there was a missing part, we designed primers to fix it.

We also continued with building our models. We continued working on the feed forward loops and made the first version of our network building algorithm.

In this week, we mainly focused on tutorials. As the input RNA and AND gate inserts were incorrect, we also did some troubleshooting and repeating of these experiments. The overlap extension PCR was repeated for the AND gate, after which we also designed new primers. We performed a golden gate cloning experiment of the first toehold and finished the plasmid construct for it.

By now, we understood SwitchMi Designer and started working on improving the software tool. We added some code to obtain a figure of the secondary structure of the toehold switches.

After designing them, we were ready to construct the DNAzyme components and the TMSD parts. For the input plasmids of the toehold gates a golden gate, transformation, and a colony PCR was preformed. However, in the sequencing results, we saw that the promoters were incorrect. A PCR was preformed on the AND and OR gate plasmids, after which they were cloned. After the literature study, we also started with the design of the sponges strategy with BioBricks. The marionette strain was also transformed with our first toehold switch plasmid. By adding the missing tet, the trigger plasmid was also established for the first toehold switch.

For the in silico toehold design, we continued working on the code to make the figures with the secondary structure. We also used synthetic data on our first feed forward models, which could form the basis for experimental design later on. By using adjacency matrices, we build a new version of the network building algorithm.

This week, we performed our first test of the DNAzymes. We also finished the production of the TMSD parts. For our amplification method, we started in the lab testing NASBA with QUBIT. Qubit was bound to the primers which caused the signal to be too high. Therefore, we decided to test NASBA in steps. Even though, we obtained the correct AND gate, we needed to do some troubleshooting with the OR gate sequence. We also redesigned the sponge strategy and ordered ologo’s for these experiments. After repeatedly transforming the marionette strain, we still found our plates to be overgrown. When researching, we found a problem in the resistance of the strain. Since a different resistance gene was used on the trigger plasmids, we now first transformed this into marionette.

After HP talked to some stakeholders, we decided to focus on diagnosing relapsing-remitting multiple sclerosis. Therefore, we could start search for data on this disease. We also started on modelling the TMSD to predict the optimal parameters to be used in the lab. By expanding the network building algorithm, we created an algorithm that can generate all possible combinations of logic gate circuits.

As the DNAzyme did not work, we continued troubleshooting them. We had to design and order new primers for LacZ. As NASBA needed to be tested in steps, we also designed new constructs for this experiment. The contructs of TMSD were made, so the transcription of them was tested in vitro. This production of the AND, and OR gates, including both the LIRA and toehold based gates, continues with PCR and cloning, and the produciton of the corresponding trigger plasmids needed some more troublehooting and designing. The sponges were made with golden gate cloning, after which a transformation and colony PCR were preformed. As the marionette strain was found to have a resistance problem, we recloned the toehold switch construct to a kanamycin Resistance plasmid. The marionette strain with trigger was made competent with an electroshock and the new toehold switch plasmid was added.

The data found for relapsing remitting MS was researched and processed according to the paper it was from. We continued with the model for the TMSD. After checking the conditions, we also started with making an optimisation algorithm for the circuit design.

This week, we were mainly busy with troubleshooting the performed experiments. We continued our experiments with testing and troubleshooting the DNAzymes. This is also the case for the sponge constructs. The construction of the paper-based constructs were also wrong so we continued with troubleshooting these for the next four weeks. The aptamer for the TMSD was optimised and the final NASBA constructs were ordered. The production of the different toehold logic gates also needed troubleshooting and redesign for most constructs.

In our dry lab, we researched the possibility to add a model predicting the ON/OFF ratios of the toehold switches, continued with the TMSD model and generated the first version of the optimisation algorithm for the circuit design.

For the DNAzymes we tested different buffers and designed alternatives DNAzyme options. After successful cloning of the miRNA, we still had to redesign the golden gate cloning for the sponges. We were also still in the process of troubleshooting the paper-based constructs. This week, we optimised the IVT for the TMSD. For the first step of NASBA, we made a positive control with PCR, and tried reverse transcription, which gave some unclear results indicating some problems. In the toehold AND gates new RBS sequences were designed to replace the old once, and PCR and cloning was performed for this. For our first toehold switch, we prepared buffers and M9 medium and our first plate reader experiment was performed. However, no signal was seen.

This week, we continued with the miRNA expression data and filtered it according to the paper the data is from. The TMSD model was also not finished, so we continued with that. A second version of the optimisation algorithm was made by using pyPESTO and latin hypercube sampling.

For the RNA sponges, we performed a golden gate method with 4 templates instead of 9. We were still troubleshooting the paper-based constructs. Because we wanted to use a different aptamer, we also redesign the other TMSD parts. The reverse transcription and RTPCR of NASBA were repeated, but the results were still unclear. In addition, IVT was tested, but did not yet work

For the TMSD model, we found the optimal parameters to be used in the lab. We also optimised the model for a more complex system. As the optimisation algorithm was finished, we could find the optimal circuit design.

In this week, we repeated the golden gate cloning for the sponges but now with different ratios of the components. We were also still in the process of troubleshooting the paper-based constructs. For the TMSD reaction the in vitro transcription was tested and for NASBA the step of reverse transcription was tested with an in vitro transcription reaction. However, for the later the reaction did not perform properly. To produce a postive control for the toehold tests, a PCR was performed on only and RBS with GFP. By this time we were also able to do some tests in a PURE system and do some test with the kit provided by Promega.

The miRNA that will bind to the sponges were synthesised. For the sponges, we tried to find the best template this week. We continued with troubleshooting the DNAzymes. For the PURExpress system we experimented with lower reaction volumes. As the promega kit was unable to give us a proper output, we cloned a chromoprotein as an alternative construct and tested different variables for Promega. To troubleshoot NASBA we used different primers. However, the reverse transcription only worked when we used a lower concentration of only the reverse primer. We also amplified the LIRA AND gate after which it was cloned. A plate reader experiment was also done for the toehold switch with different promoters (p105 and p100) after a positive control was constructed.

In the dry-lab, we continued working on our models. We also performed a random forest model to obtain the most important miRNA in MS.

As the golden gate reactions for the sponges did not work, we redesigned the experiment to make a one pot reaction. After some time of troubleshooting the DNAzymes, we found that an old bottle of H₂O₂ was the problem. The NASBA step of reverse transcription was repeated for an MS related miRNA. After the promoters, Psyn and Ptet, arrived we were able to perform a golden gate and finish the input plasmid and the gates. We repeated the golden gate to change the RBS and we added a hairpin to the AND gates. With a plate reader experiment, we checked the leakiness of the correct gates. Using a PCR, we remove extra bases from the trigger. We also fixed to Ptet promoter on the positive control plasmid of the toehold switch. The transformation for the production of the chromoprotein for Promega testing was not yet successful.

To find miRNA to distinguish between MS and healthy control, we used another classification technique (histgram gradient based classification). However, this was not used further. The code was also improved to enforce a better miRNA annealing to the predicted toehold switches. We also build a model that was able to predict the TMSD reaction in the lab. We compared the model output with the data obtained in the lab. The boolean network for the prediction of the AND gate was also converted to a continuous model.

This week the first golden gate one pot reaction was performed on the sponges with a two step protocol. As the DNazymes of the paper don’t work due to hybridisation of H3 and H4, new DNAzymes were designed such that it anneals to the linear domain of the toehold switch. For the amplification of the MS miRNA, an in vitro transcription on the RT PCR product was performed. A colony PCR shows all AND gate and input plasmids are correct. By amplifying the plasmid backbone with GFP, we reduced the amount of constructs in the golden gate reaction. This was done to troubleshoot the wrong gates. We also build a plasmid for a trigger RNA and toehold switch plasmid for a second toehold. Primers to make toehold switches targeting an MS miRNA and to add a hairpin to the trigger plasmid were also designed. We repeated production of chromoprotein for in vitro testing, but again did not work, so we designed new priemrs for gibsom assembly for this

This week, we designed toehold switches targeting MS miRNAs in silico. We continued with the analysis of the data obtained in the lab with the model. The model predicting logic gates switches was also made and simulated.

We unfortunately decided to stop working on the sponges, due one of the team members getting ill. In this week, we finished the production of all different RBS an hairpin sequences for the AND gates. We also troubleshooted the modular input RNA and reconstructed the toehold gates. With overlap extension PCR we obtained a plasmid for the toehold switches detecting an MS-miRNA. Since new PURE kits arrived, we could perform a pretreatment for the paper test, including cutting out some discs, and drying these. We also tested the first paper discs using the positive control plasmids. During this week, we also performed overlap extension PCR to construct all 3 toehold switches targeting the MS miRNA, and cloned the chromoprotein.

We found an MS-specific miRNA combination with our models and continued working on the TMSD models optimising the parameters. We also fitted the data obtained in the lab of the logic gate circuits to the model predictions.

The construction of RNA switches, trigger and DNAzymes parts were finalised. An in vitro transcription was performed for TMSD and the RNA folding was optimised. A RT transcription was performed on the MS miRNA of which an PCR was performed. In addition, the final IVT test for NASBA was performed, and larger reaction volumes were tested for this. For the toehold AND gates, double transformations of both trigger and circuit plasmids of marionette strain were performed, of which the colonies were used for plate reader experiments.In addition. Overlap extension PCR was performed to obtain the MS trigger plasmid after which an PCR was performed. To make the toehold switch plasmid for the in vitro experiments, a golden gate was used. The final toehold plasmids, including one AND gate, were also tested in vitro (PURExpress) for a qualitative test.

To take other disease into account in our test, we searched for miRNA expression data for mimic diseases. We also performed a parameter space exploration for TMSD.

In one of the last lab weeks, we tested the RNA DNAzymes in tubes and tested out first toehold switch on paper. For the TMSD reaction, we tested the optimisation of the folding, the CD activation and the B inhibition. We also tested the limit of the detection of our MS miRNA for RT. We also tried to replace our inducible promoter for the trigger plasmids with an constitutive promoter, as a control for the failing in vivo experiments. For the AND gate, we changed the copy number of the plasmid and the degradation tag after GFP. Plate reader experiments were performed on the gates. The trigger RNA plasmid for the LIRA gates were cloned. We also finished the trigger constructs and transformed everything in the marionette strain with multiple methods for the MS toehold. As there were no colonies visible, we decided to stop the in vivo experiments.

In our final lab week, we did the final tests with the DNAzyme. We also tried to quantify the toehold output and performed an accuracy essay for the toehold switch. In addition, the majority of our team was working on the wiki from this point onward.

The data obtained in the lab on the AND gate was fitted to the data obtained from our toehold circuit model.

While most wiki pages were finalised in this week, we still found some time to use our circuit model to gain more insights into the leakiness of the toehold switches and circuits.