“Science is the art of systematizing and documenting observations.” - Louis Pasteur

Biological Team Lab Journal

On this section, we will outline the series of steps that guided the biological team in optimizing experiments and validating the model. Furthermore, we will elucidate the process of developing the protocol for origin of replication replacement. For comprehensive and detailed experiment protocols, please refer to the experiments page following the 🔎 sign.

3.4- Receiving puc19:

receiving E-coli with puc19 from Uri Gophna's lab. The E-coli cells that carry the plasmid were given to us on an LB plate containing ampicillin, an antibiotic to which the plasmid carries resistance.

4.4- Biology team meeting:

Planning the preliminary experiments. The puc19 we received did not carry a reporter gene, so we decided to use the iGEM part collection and perform a ligation between the segment of their reporter gene and Puc19 plasmid.

9.4- Starting experiments:

Transformation of iGEM parts plasmids with reporter genes to E-coli dh5 alpha cells that maximize transformation efficiency. The parts were selected according to the proteins they encode and are detailed in the chart below:

Part number Name ORF part protein
1 BBa_K876070 BBa_K081012 GFP
2 BBa_K1692032 BBa_K592009 amilCP blue chromoprotein
3 BBa_K1033900 BBa_K1033901 meffBlue, blue chromoprotein
4 BBa_K1033903 BBa_K1033905 tsPurple, purple chromoprotein
5 BBa_K1033907 BBa_K1033909 fwYellow, yellow chromoprotein
6 BBa_K1033914 BBa_K1033916 amajLime, yellow-green chromoprotein
7 BBa_K1033923 BBa_K1033925 spisPink, pink chromoprotein
8 BBa_K1033917 BBa_K1033919 gfasPurple, purple chromoprotein
9 BBa_K1033926 BBa_K1033927 asPink, pink chromoprotein
10 BBa_K1033928 BBa_K864401 aeBlue, blue chromoprotein
11 BBa_K1073024 BBa_K592010 amilGFP, yellow chromoprotein
12 BBa_K1073026 BBa_K592011 cjBlue, green chromoprotein
13 BBa_K1033282 BBa_K1033930 amilCP, blue/purple chromoprotein


The plasmids were introduced into the competent bacteria through a chemical transformation process🔎. The bacteria were seeded on plates given by Uri Gopna’s lab, containing LB with chloramphenicol, to which the plasmids carry resistance.

10.4- Testing the transformation:

The transformation was successful in 10 out of 13 parts, in which the bacteria were grown on the plates containing the chloramphenicol antibiotics. In part of the plates, we were able to see clearly an expression of color.

pies

16.4- Choosing best parts:

Our next goal is to connect puc19 to a reporter gene found in iGEM parts. For this purpose, we selected 3 parts that expressed the strongest colors (part 2, part 8 and part 9), and patched them to new plates🔎. In addition, we prepared starters🔎 for the bacteria.

17.4- Restriction digestion and igation:

The next morning the patches of the parts showed strong colors. The BBa_K592009 liquid LB that encodes to blue chromoprotein also showed color.

laboratory work 2 laboratory work 3


In order to isolate the plasmid, we used NucleoSpin Plasmid EasyPure, Mini kit for plasmid DNA purification🔎. Then, a NanoDrop machine🔎 was used to read DNA concentration and purity. With this information, we continued the restriction digestion and ligation using the enzymes EcorI and PstI🔎. These restriction enzymes have cutting sites around the chromoprotein found in the 3 different parts. In addition, they are located within the multiple cloning site of Puc19, so that after the ligation process, the fragment of the protein can be appropriately inserted into the puc19 plasmid.

plasmids


A DNA ligation reaction was performed using T4 DNA ligase🔎 incubated with DNA in the ratio of 1.8:1 between the insert and the vector fragment . The ligation products were transformed into E-coli bacteria and seeded on LB plates containing ampicillin.

18.4- Testing the ligation and designing sequences:

Both the ligation and transformation worked. At this point, we have created 3 different pUC19 plasmids all with an identical origin of replication, but with a different reporter gene (different chromoprotein). In addition, we decided that the mutations for the ORI in the next experiments will be done by the same technique of restriction digestion and ligation, hence we designed primers that will surround the ORI fragment and will contain two restriction enzyme sites that don’t appear in our plasmid. In addition, we designed different ORIs with mutation in RNAi or RNAp that contain the same restriction enzyme sites.

ori

24.4- Growth experiment:

Checking the growth of the E-coli containing different plasmids- the original plasmids from the part collection and puc19 with the chromoproteins (after the ligation). The growth is measured by the optical density (OD) at 600 nm🔎. By looking at the slope of the growth curve of the different bacteria, no significant differences in growth were found between the different plasmids.

plas


14.5- creating glycerol stocks:

Bacterial glycerol stocks🔎 are used for long-term storage of plasmids. They ensure that the desired plasmid DNA is readily available in the desired bacterial strain without the need for additional cells. Glycerol stocks preserve bacteria for extended periods by protecting cell membranes and maintaining cell viability.

15.5- Chromoprotein Plate Reader Wavelength Calibration

We used a plate reader to investigate whether we could measure the copy number of different plasmids with chromoproteins as reporter genes, similar to fluorescent markers. We used the plate reader, assigning different excitation and emission values to distinguish between the plasmids with chromoproteins and the negative control. Additionally, we scanned for absorbance values between wavelengths of 350-700 nm.

Unfortunately, we couldn't clearly differentiate between the control and the plasmids containing the target gene. This led us to realize that this method is not suitable for measuring the copy number of plasmids without a fluorescent protein. Therefore, we decided to use the qPCR method exclusively.

16.5- Growth experiment with plate reader:

To validate the findings of our growth experiment conducted with a spectrophotometer, we replicated the study using a plate reader 🔎. The growth rates of the bacteria containing the different plasmids were found to be similar during this experiment as well.

graph2 colordots

29.5- First use of qPCR:

We decided to use a quantitative real-time PCR (qPCR) technique to determine the PCN of the plasmids. We designed specific primers🔎 that targeted both the bacterial chromosome and the plasmid. To calculate the PCN, we will divide the number of plasmids by the number of bacteria.The approach we decided to use is colony qPCR🔎. Colony qPCR is a method that directly analyzes DNA from bacterial colonies. It avoids DNA isolation, preventing DNA loss. In colony qPCR, bacterial colonies are placed in a qPCR machine with high temperature to lyse the cells and release DNA. This approach saves time, reduces DNA degradation and contamination risks, and enables real-time quantification of the target DNA.


6.6- Creating standard curve:

Creating a standard curve🔎 in qPCR is important for quantification and accurate measurement of target DNA. The standard curve allows for the determination of the relationship between the cycle threshold (Ct) values and the initial concentration of the target DNA.

The generated calibration curve had a low R2 value, indicating poor linearity. Moreover, the obtained Ct values were significantly low in this experiment, suggesting the need for a further dilution of the DNA sample solution.

20.6- Repeating standard curve experiment:

We repeated the qPCR experiment one more time. This time we performed colony qPCR- we selected a single colony, mixed it with 10 μl DDW and then started with decimal dilution. The Ct values of the plasmids were too low, suggesting that the plasmid amount is very high in the colony in comparison to the bacterial chromosome. We concluded that for the plasmid’s qPCR reaction we must make more dilutions than from the bacterial reaction. The standard curve of the bacterial chromosome was satisfying.

graph4


Based on the slope of this curve, we calculated the amplification efficiency of the chromosomal gene, which we can use later to calculate the different plasmids copy number.

$$Ec = 10^{\frac{-10}{-3.147}}=2.078$$

$$Ec{(\%)} =(2.078-1)*100=107.8 $$

25.06- Preparing vector and inserts for cloning new plasmids:

In our project we are trying to create mutations in the RNAp or RNAi promoters and see how it affects the plasmid copy number. To achieve this, we are replacing the origin of replication with a mutated version. We decided to do it using restriction digestion and ligation.

We received the PCR🔎 primers for creating a plasmid backbone without the Ori, and with cutting sites for restriction enzymes at the ends. We also got the Ori segments ordered with matching restriction sites at their ends. For this first iteration of replacing the Ori and measuring the new plasmid copy number, we used 8 sequences from the article of Rouches et al: ”A plasmid system with tunable copy number”.

26.06- restriction and ligation:

Using an agarose gel electrophoresis🔎 , we checked if the PCR sample contained DNA of the desired size. We observed that the bands were about 3000 bp in size, which was consistent with our expectations of the plasmid backbone.

prc


Then we proceeded to restriction digestion and ligation to connect the new ori fragments with the plasmid backbone that contains the same cutting sites at the ends. We incubated the DNA in the ratio of 3:1 between the insert and the vector fragment . The ligation products were transformed into E-coli bacteria and seeded on LB plates containing ampicillin.

27.06- Trying different ratios:

The process did not work - no colonies were obtained in the plates transformed into the new plasmids. we thought of several possible reasons:

  1. We did not inactivate the restriction enzymes - after checking, BamHI does not have an inactivation temperature.
  2. There is not enough insert in relation to the backbone
  3. The ligation process is not finished.

We decided to make a new transformation with the ligation reaction that stayed overnight in 4 °C . In addition, we tried to increase the amount of the insert relative to the backbone and to make ligation with larger amounts of the backbone.

28.06- Checking the plates

The next day, there were almost no colonies on the plates as the process failed again. Colonies also grew in the plate of the negative control in which we transformed the backbone PCR product without an insert. That is, it is possible that during the process the original Ori of the plasmid, that remained after the PCR process in very low quantities, is closed together with the backbone, and these are the colonies that we see growing.


2.07- Thinking about how to solve the problems:

After consulting with the committee member Prof. Itai Benhar, he suggested that we use an additional restriction enzyme that cuts in the middle of the ORI so that no original ORI remains and the negative control will not contain any colony. After checking, we found that the AlwNI enzyme is suitable for this idea because it cuts the plasmid only once in the ORI region and in addition it has an inactivation temperature. Before trying the new idea, we performed a preliminary experiment to check that the three restriction enzymes - BamHI, SalI and AlwNI work properly. For this experiment we took an original puc19 plasmid and each time cut it with a different pair of plasmids. We then ran the products in an agarose gel and saw that the three restriction enzymes were indeed working.

3.07- Adding the new step:

We followed the steps of the reaction with the additional step- we incubated the vector with the three restriction enzymes BamHI, SalI and AlwNI for 1 hour in 37 °C, we inactivated AlwNI for 20 minutes in 80 °C, and then performed ligation to the cut inserts using T4 DNA ligase. We incubated the DNA in the ratio of 2:1 between the insert and the vector fragment . The ligation products were transformed into E-coli bacteria and seeded on LB plates containing ampicillin. We kept the remaining ligation product in 12 °C overnight in case the reaction doesn't work since we found it helps to complete the ligation process.

4.07- Trying more options:

On the next day, we observed that no colonies had grown on the plates. We were somewhat pleased to observe that there were no colonies present on the negative control, indicating that the extra step recommended by Professor Benhar was effective.

plates


This day we tried to do a transformation to the ligation product that stayed overnight at 12 °C. We also tried to increase the amount of two inserts in relation to the vector, to ratios of 5:1 and 10:1.

5.07- Observing the results:

The ligation of RNAi1, using a 10:1 ratio, appeared to be successful, resulting in the development of 10 colonies. After successfully ligating RNAi1 with a 10:1 ratio, we proceeded to create patches from these colonies and continued with the process of increasing the amount for other inserts.

6.07- Continue the attemptions:

The patches of RNAi1 didn’t show color of the chromoprotein, maybe because this Ori is supposed to have a low copy number. We kept the plate of the patches in the refrigerator and continued to watch it. The plates that we seeded the previous day didn’t grow.

oneplate

9.07- New ideas to continue

Last week we saw that the restriction enzymes work properly and we were able to reduce the negative control to zero with the AlwNI enzyme. However, ligation and transformation only worked once with low efficiency (in the RNAi1 experiment at a ratio of 10:1). When we tried to repeat this success it didn't work, hence in the coming week we decided to try some new things:

  1. To increase the amount of the ligation product which we transform into bacteria.
  2. Mix the inserts - create a mix of all the ORI sequences to increase the chances that we will get all the mutation options once the reaction is successful.
  3. If this does not work, we will try to increase the efficiency of the transformation with the help of electrophoresis.

In addition, we decided that if all these attempts fail, we will think about changing the approach for the experiment to replace the ORi of the plasmid.

We made ORI mix by adding 5 ÎĽl from each of the ORI sequences to a new eppendorf tube.Then we performed ligation with a higher amount of the plasmid backbone and transformed 10 ÎĽl from the ligation product to the bacteria.

10.07 - More attemptions:

Once more, no colonies were observed on the plates. Undeterred, we carried on with our experiments. On this particular day, we attempted to use reduced quantities of DNA for both the insert and the vector, maintaining a 10:1 ratio. Furthermore, we performed a transformation using the ligation product from the previous day that had been refrigerated. All transformations on this day were done with the help of electric transformation🔎 in order to improve efficiency.

11.07- Watching the results:

Unfortunately, the attempts and efforts did not work and again we did not get colonies in the different plates.

manyplates

13.07 - Changing the method of the cloning:

After consulting with the members of the laboratory of committee member Prof. Martin kupiec, we realized that the process does not work properly for us since restriction enzymes do not cut optimally at the ends of the DNA segment. Therefore, we decided to change our approach and perform the cloning using Nebuilder assembly. This technique works by joining DNA fragments by using their natural ability to recombine. For this process, we need to have overlap sequences that complement each other between the new ori fragment and the plasmid backbone. Since we already had eight ori sequences that we wanted to insert into the plasmid, we ordered primers that would extend them on both sides and create the homology to the backbone of the plasmid.

18.07 - Nebuilder Assembly:

The primers arrived and we performed PCR for the ORI segments in order to extend them at the ends and create homology to the backbone of the plasmid. After that, we performed Nebuilder assembly🔎 according to the quantities required in the protocol and performed transformation with different volumes of the product- 2, 5 and 10 µl.

20.7 - Agarose gel electrophoresis:

New PCR reaction on the plasmid to get the backbone of the plasmid (without the ori).

19.7- PCR:

We performed agarose gel electrophoresis on the PCR products. there wasn’t a strong band.

23.7 - PCR with Taq Q5 and agarose gel electrophoresis;

Since the previous PCR didn’t have a strong band in the agarose gel, we decided to perform PCR with Taq Q5.Now we perform agarose gel electrophoresis and have a strong band in the gel. From comparison to the DNA ladder we conclude that the concentration of the PCR product is 20 ng/μl. After the quantification of DNA concentration, we redo the NEBuilder reaction with a ratio of 1:2 between the vector and the insert. In addition, we added to the transformation process a step of adding SOC- a rich bacterial medium that aids in the

24.7- More attempts:

The colonies did not grow. We tried to do transformation with all the amount of the NEbuilder reaction (20 ÎĽl) and to increase the ratio of the vector and the insert to 1:10.

25.7- DNA purification:

Since the plates were empty again, we decided to try to purify the PCR products before performing the NEbuilder reaction in order to optimize it. We tried again the original ratio of 1:2 between the vector and the insert, and tried to use several amounts of the product for the transformation.

26.7- Thinking of the next approaches:

Among the different plates we seeded, we noticed that the plate that underwent a transformation of 2 ÎĽl showed growth of colonies. From this we conclude that 2 ÎĽl is the most effective amount for the transformation process. We patched the colonies that grew and monitored them.

Since the efficiency was low, we started thinking about other ways to change the origin of replication of the plasmid. After consulting with our expert’s committee we thought of trying the pGEM vector systems. These are convenient systems to clone PCR products. The pGEM vector is a high-efficiency cloning vector which contains multiple cloning sites. This could solve the problem we had in the restriction process, when the cut sites were at the edges.

In addition, we decided to order a new sequence of RNAp1 that we designed using the Nebuilder assembly tool. This sequence contains homology segments with the plasmid’s backbone.

31.7- Trying the pGEM vector system:

We cleaned the PCR product of the plasmid backbone and ligated to pGEM®-T Easy Vector Systems using T4 ligase. The ligation product was transformed to the bacteria.


1.8 - Keep trying:

The transformation of the pGEM vector succeeded. We created starters of their colonies and tried once again the NEbuilder method with the purified PCR products.

2.8 - New ORI sequence:

The new ORI sequence with the proper homology segments arrived. We diluted it only with 10 ÎĽl so it will be more concentrated and performed the Nebuilder assembly according to the original protocol.

3.8 - Success:

Colonies were grown on the NEB reaction plates. After a gel run test we saw that the process was successful and the colonies contain plasmids. We finally managed to find and specify the protocol for replacing the origin of replication!

purpule

7.8 - sending the new colonies for sequencing:

To confirm the presence of mutations in the new colonies, we submitted them to the sequencing unit at Tel-Aviv University. Sanger sequencing was performed using custom-designed primers.

9.8 - sequencing results:

We successfully detected the mutation in RNAp1. Furthermore, we uncovered another mutant with a distinct modified nucleotide. This suggests that the introduction of new sequences can sometimes lead to unexpected mutations. We called the new mutants mut1 and mut2 and made patches from them.

copynumberplate


Now that we saw that the NEB process was working properly, we ordered the rest of the sequences with the homologous segments.

14.8 - Nebuilder with the new sequences:

The rest of the new sequences arrived and we inserted them into the backbone of the plasmid with NEbuilder assembly.

15.8 - Checking the plates:

new colonies grew on the plates and we made from them starters. We realized that we came up with the method for replacing the origin of replication and that the last time we succeeded was not accidental.

platewithbubles

16.8- differences in color:

Since the plasmid we used carried the pink chromoprotein, we were happy to see that the different colonies have different shades of the color. The differences were also notable in the starters. We could estimate from the color if the colony has a high or low copy number plasmid.

pipets

17.8- Sending samples for sequencing:

We sent the products of the last NEB reaction to sequencing.

27.8- improving puc19 standard curve:

Since in the previous calibration experiment we got low Ct values for the plasmid, we performed another qPCR experiment in order to improve our standard curve, this time we diluted the colonies 200 times more. Unfortunately, in this experiment the Ct values were again too low and did not allow creating a reliable calibration curve.

In addition, we decided to make plasmids of the different mutant origin of replication with YFP - a fluorescent reporter gene. We did it using restriction digestion and ligation.

We sent the products of the last NEB reaction to sequencing.

30.8- qPCR for calculating plasmid copy number:

We conducted qPCR to determine the plasmid copy number in the new mutants. For each mutant, we carried out triplicate qPCR reactions, targeting both the plasmid and the chromosomal gene.

31.8- New Nebuilder assembly reaction:

We received new ORI sequences for our second iteration. We diluted them and performed the assembly reaction with the plasmid’s backbone.


1.9- The assembly succeed:

The assembly reaction was successful and we could watch colonies of the new mutants. We made patches and starters for next experiments.

amp

4.9 - Checking sequences + qPCR:

Plasmid purification from starters made from colonies of the nebuilder product plates. The plasmids were sent to sequence in order to detect new mutants.

We conducted qPCR in order to improve the standard curve of the plasmid. This time the Ct values were satisfying and were plot against dilutions.

puc19graph

6.9 - qPCR:

qPCR for 5 mutants found in order to measure their copy number. triplicates were made for each mutant.

7.9 - Checking sequences:

Since we didn’t find all the mutants in the previous sequencing session, we purified plasmids from other colonies and sent them to sequencing.

10.9 -qPCR:

Repeating qPCR of 3 mutants with different dilutions (100,1000) in order to get signals in the optimal range of the qPCR machine (20<Ct value <30)

11.9 -qPCR

Because the previous dilutions weren’t enough to get the signals in the optimal range, we made 6 serial dilutions of the same colony (10-4 to 10-9), and performed another qPCR. Fortunately, the dilution of 10-6 was in the right range of the Ct signal as seen in the graph.

bargraph1


After we found the right dilution we performed qPCR on other mutants and calculated their copy number.

12.9 - growth experiment + qPCR:

We conducted an experiment to assess the growth rates of the various mutants. We initiated the experiment by measuring the optical density (OD) of each mutant culture. then an equal amount of each

growgraph


We also performed qPCR measurements of the rest of the mutants in order to calculate their copy number.

13.9 - Checking sequences:

Sending more plasmids to sequencing in order to detect all the mutants from the last NEBuilder reaction.

19.9 - Nebuilder reaction + transformation:

We performed another NEBuilder reaction in order to get more sequences with new mutations and transformed E.coli to overnight incubation.

20.9 - qPCR-

We performed qPCR on other 7 mutants and calculated their copy number.

21.9 - Checking sequences:

Plasmid purification from starters made from colonies of the last NEBuilder process and sequencing.

27.9 - qPCR-

qPCR on 7 found mutants and copy number calculating.

2.10 - qPCR-

qPCR on 3 more mutants and copy number calculating.


January

  • Introduction with the steering team: Prof. Avigdor Eldar, Prof. Martin Kupiec, Prof. Itay Benhar, Prof. Ori gofna.
  • Starting to think of ideas.
  • Consultation in zoom with Dr. Shay Tal an expert in plasmids research.

February

  • Literature review and Data search: After choosing and developing our project idea, we did an extensive literature search to find more information about the Col-E1 replication system, existing tools (none were found) and databases that will help us build our model.
  • Presentation of the idea at the Israeli biology conference “Ilanit conference”.

March

  • Development of project plan.
  • Data analysis: we found an article [ref] that contained a library of pUC19 plasmids, each containing a different promoter sequence for either priming RNA (RNAp) or inhibitory RNA (RNAi) whereas the variations are only in the promoter regions of either RNAp or RNAi and decided to use this data to design our model. We read through this article and understanded all its aspects. We visualized the copy number distributions, recreated and validated the Promoter Strength feature that appeared in the original data.
  • Consultation in zoom with Prof. Manel Camps an expert in plasmids research

April

Week1
  • Design basic sequence based features: such as One hot encoding features and Nucleotide counts features.
  • Organization of lab space, starting of the experiments: transformation of chromoprotein parts to E. coli.
Week2
  • Correlations: in order to further understand the data, we calculated the correlation scores (pearson and spearman), as well as the Mutual information coefficient between all the features and the Copy number value. PVs were calculated as well
  • Choosing the optimal parts to our experiments and replacing the backbone to pUC19.
Week3
  • Basic model: we built a basic Linear model, for RNAi and RNAp separately.
  • Examining the parts in the new backbone and starting growth experiments.
Week4
  • Evaluation metrics and visualization: we decided to evaluate our model using spearman, pearson, MAE and R-squared score on a test set. We also created graphs of actual Copy number values Vs. predicted values, and feature importance graphs.

May

Week1
  • Advanced Model: since we weren't satisfied with the linear model results, we looked for more advanced models, and built XGBoost models for RNAi and RNAp separately.
Week2
  • Shared model: we tried to build a model that takes under consideration both RNAp and RNAi mutations, but our data didn’t fit this model and we decided to leave it.
Week3
  • Design advanced features: in order to improve the models performance we added more advanced features such as: PSSM , motifs scores and PV from different databases , denovo motifs using HOMER
  • Chromoprotein Plate Reader Wavelength Calibration, growth experiment with plate reader.
Week4
  • Hyperparameters tuning: we’ve implemented basic hyperparameters tuning using grid search for our models.
  • First use of qPCR.

June

Week1
  • Feature selection: we’ve implemented basic feature selection metric (sequential forward feature selection) for our models.
  • Creating standard curves.
Week2
  • Models and features analysis: we’ve performed model analysis regarding the model structure (XGBoost complexity analysis: model performance vs. number of trees and the maximal depth of the trees And regarding the model features: The models were trained 1000 times and the 40 features of each model with the highest features importance score on average were intercrossed. The models were trained for 5000 iterations, when the data was randomly splitted each time and counted the number of times each feature was used in the models out of 5000 iterations. The feature importance and average R-squared score were calculated as well.
  • Repeating standard curve experiments.
  • Ort school visit in the university.
Week3
  • Hyperparameters tuning: we've upgrade the model hyper parameters optimization with a library named “Optuna”, which uses bayesian search in order to find the optimal hyper parameters.
  • Preparing vector and inserts for cloning new plasmids.
Week4
  • Feature selection: we’ve upgraded the model feature selection. We calculated the mutual information (MI) between each feature and the copy number, selecting those with MI values greater than the mean MI as potential candidates. Subsequently, we assessed the MI between the remaining features, identifying pairs of features exhibiting strong correlations within the top percentile. For each pair of correlated features, we chose the one displaying higher MI with the copy number, ensuring the retention of the most informative attributes. Last, we used a library named “eBorute” to further select the model's important features.
  • Isragem competition

July

Week1
  • Exploratory Data Analysis: we’ve visualized the chosen features and created graphs that display the features distribution and the features values vs. the copy number.
  • Consulting with the committee member Prof. Itai Benhar to solve problems in the cloning process.
  • Adding the new step to replacement of the ORI protocol, restriction with AlwNI.
  • More attempts to replace the ORI.
Week2
  • Copy number prediction: we used our models to predict the copy number for all the possible sequences
  • Brainstorming to overcome the cloning problem.
  • Consulting with the members of the laboratory of committee member Prof. Martin kupiec, and changing the method of the cloning to NEBuilder.
  • Consultation in zoom with Miles Rouches an expert in plasmids research
Week3
  • Adjust the pipeline: we changed the feature extraction code to calculate only the selected features for all possible sequences,in order to enhance model efficiency.
  • Update the copy number fitting of the article based on biological results from our lab, in order to improve the model predictions correlations to web lab results.
  • Backbone plasmid purification for NEbuilder assembly.
  • Tuller’s lab mini-conference
Week4
  • SHAP explainability analysis: We used a library called “SHAP” (explainable AI with Shapley values) to evaluate the features that were chosen and their effect on the decision process of the model. In this way, we can determine which features are critical and which values of each feature contribute to a high copy number.
  • Using the updated model prediction for our user friendly online software tool.
  • NEbuilder assembly, thinking of the next approaches: pGEM vector system and designe using the Nebuilder assembly tool. Trying the pGEM vector system.

August

Week1
  • Conversation with the article’s author: we consulted with the author of the article, the person who created the variantes library in order to deeply understand the methods of the article, how they generate the data, consulted about our biological methods and more.
  • Another try of NEbuilder with the right homology sequences - success.
  • Sending the new colonies for sequencing to identify the mutations.
Week2
  • RNA structure features: after consulting with the article’s author, and understanding an important fact about the RNAi promoter variants, we added more than 10,000 new features of the RNAp structure changes based on the RNAi promoter sequence.
  • Understand the need: we’ve consulted with biologists to understand the important features we need to include in our tool.
  • Sequencing results, we detect RNAp1 and other random mutation.
  • Entrepreneurship day led by Yair Sakov.
Week3
  • Voting model: we wanted to further enhance the model performance, so we decided to examine additional model’s types. We analyzed their performance, and chose to design a voting model where the final prediction is the average of the predictions of the best models.
  • Nebuilder with the new sequences, sending samples for sequencing, detection of other mutant.
Week4
  • Model confidence:To assess the accuracy of our regression model's predictions, we used quantile regression to calculate confidence intervals for the upper and lower bounds of the predictions. Quantile regression is a type of regression that models the quantiles of a response variable, rather than the mean. This allows us to quantify the uncertainty in our predictions and to be more confident in our results.
  • Improving puc19 standard curve,qPCR for calculating plasmid copy number, New Nebuilder assembly reaction.
  • Meet up with the Technion iGEM team.

September

Week1
  • Engineering cycle: we updated and fitted the copy number values of the article based on biological results from our lab, in order to improve the model predictions correlations to wet lab results
  • Checking sequences of the latest assembly, and qPCR experiments to calculate plasmid copy number.
Week2
  • Building the tool: we started to build the tool, to give users the option of achieving desired copy number, given any col-E1 plasmid.
  • Visualization: we re-created all the figures using python package “Plotly” [ref] to achieve interactive figures to display on our wiki.
  • Optimize the qPCR experiments to get Ct values in the right range of signals, and repeating of qPCR experiments in the right dilutions.
  • Growth experiment in the plate reader. Synbio conference.
Week3
  • Clustering: To identify groups of promoters with similar chimera scores, we used a variety of clustering algorithms, including HDB and SP, which identified 17 and 46 clusters, respectively. Analysis of the promoters in each cluster revealed that one cluster contained sequences exclusively from the Yersiniaceae family, suggesting that a ColE1 system may be present in other organisms.
  • New Nebuilder reaction.
  • Performed qPCR on other 7 mutants and calculated their copy number
  • Visit at alagene
Week4
  • Designing the tool: we consulted with UI experts in order to achieve a user-friendly and intuitive tool.
  • qPCR on 7 found mutants and copy number calculating

October

Week1
  • Tool output: We used the updated model prediction for our user friendly online software tool.
  • qPCR on 3 more mutants and copy number calculating.
  • Shooting the promotion video.
Week2
  • Zoom meeting with twist bio-science.
  • Zoom meeting with lonza