Promoter&Terminator Design

Considering that the target gene has to be transferred into Streptomyces, how to secure the expression of gene becomes a crucial issue. In this case, adding appropriate promoter, RBS and terminator is necessary.

We access to Genbank:KR131846.1 which contains sequence with PermE, RBS and fd terminator. We use SnapGene to add these part precisely to alkB gene, making sure the ORF is continuous. In this case, the total length of target gene becomes 1851bp.

Gene Cluster

Inspired from Luca's experiment (doi:10.1128), alkB gene usually works with our gene, that is called gene cluster according to Luca. We try to find the alkB gene cluster in Rhodococcus.

Based on Luca's experiment, We speculate that there are related genes downstream of Rhodococcus alkB.We selected a downstream fragment of 1,070 bases and compared the entire sequence with the sequence provided by Luca's experiment. The final E-value was 0, with a query cover of 74.1%. It indicates the right sequence of target gene cluster is verified.

Simulation of Cloning

To increase the probability of succession of the experiment, we use SnapGene to Simulation of Cloning. Initially, we decide the plasmids used in the experiment pUWL201 and pSF1C-A-RFP accessing from Addgene.

We use the enzyme cutting function of SnapGene. For pUWL201, we simulate EcoRI&HindIII cutting and subsequent cloning. For pSF1C-A-RFP, we do Golden-gate simulation and also subsequent cloning. In this case we have access to predicted Gel result and final sequence. You can see result at result part.

Restriction Enzyme Digestion

We use EcoRI&HindIII cutting to clone target genes to pUWL201 and Golden Gate Assembly to clone target genes to pSF1C-A-RFP.

Digestions are carried out at 37°C for 2 hours, followed by heat inactivation at 65°C for 10 minutes with enzymes, Cutsmart Buffer, ddH₂O, etc. The digested products are then purified using PCR cleanup to remove enzymes and buffer components.

Ligation Reaction

The purified digested gene fragment and plasmid are ligated using T4 DNA ligase at a 3:1 molar ratio. The ligation reaction is incubated at 16°C for 30 hours to maximize efficiency.

After ligation, Gel Electrophoresis is conducted to verify whether the ligation is successful.

Transformation & Verification

Ligation products are transformed into ET12567/pUZ8002 (E. coli) competent cells using heat shock at 42°C for 45 seconds. Transformants are selected on LB agar plates containing ampicillin (50μg/mL), Kanamycin (25μg/mL according to the manual of using ET12567) and chloromycetin (50μg/mL).

Positive clones are selected and then prepare a bacterial suspension.

Conjugation Setup

In the process of introducing the plasmid into ET12567/pUZ8002, we first culture E. coli on a shaker for 2 hours, then add the plasmid and continued culturing for 30 hours, and finally measure the OD600 value.

After the OD600 reaches the target value, we spread the E. coli used to verify gene expression on LB medium plate containing kanamycin, ampicillin, and chloramphenicol. The rest is co-cultured overnight with an equal volume of Streptomyces bacterial suspension. Subsequently, nalidixic acid is added to kill the E. coli, followed by another about 5 days of culture.

Transconjugant Selection

Streptomyces will be continuously cultured at room temperature using the method of in-test-tube plate inoculation, under the condition of being supplemented with thiostrepton (10μg/mL), until obvious phenomena are observed.

Meanwhile, we culture E. coli and Streptomyces that has not undergone plasmid introduction to prepare for subsequent tests, with the same culture conditions maintained.

CG-MS

Given that the OD of E. coli changes rapidly, we sampled 10ml from the mixed system of E. coli and C17 after one week and performed extraction with same-volume n-hexane.Meanwhile, we also sampled the control group.

We sent both the control group (without engineered bacteria) and the experimental group for testing simultaneously. The injection port temperature was set to 280°C, and the column temperature was initially set to 40°C, then increased by 20°C per minute. The experiment was conducted with the assistance of the university testing center.

Design Analysis for future

Due to multiple failures in our experiments (resulting in time constraints) and the slow growth of Streptomyces (OD600 changes at a very slow rate), we did not perform quantitative determination on the degradation by Streptomyces.

Although we no longer have time for the competition, we still hope to determine the changes in TOC (Total Organic Carbon) before and after the low molecular weight polyethylene and Streptomyces system in the future using the Total Organic Carbon method, so as to verify our final experimental results.

Data Process

To make the data readable for the model, we first convert the xlsx file into a csv format and add a Python function to verify successful data loading.

Since many entries are expressed as inequalities or ranges, we approximate them with representative numerical values. For example, for any data expressed with greater-than or less-than signs, we adjust the value by adding or subtracting 25% of the original value, and for data given as a range, we take the average of the two values. After cleaning and filtering out invalid entries, we generate the final training dataset.

Model Training

The training dataset is further loaded and standardized, then split into training and testing sets at an 8:2 ratio. It is then used to train a model based on the RandomForestClassifier, which includes a total of 100 decision trees.

The trained model is saved afterward. At the same time, the ROC curve is plotted and saved to the computer desktop.

Predictor

We design a section in the model that allows manual input of features for prediction. The input features include mulch material, thickness, the main season of mulch use in the region, as well as the average temperature and precipitation during that season.

After standardizing the input data, the model outputs its prediction along with the corresponding confidence level.

Preprocessing for Deep Learning Compatibility

The collected genomic and target gene sequence data are subjected to multiple preprocessing steps. This includes converting data into machine-readable formats, cleaning sequences to remove invalid or noisy fragments, and performing tokenization or other transformations to structure the data in a way that deep learning models (such as the Evo-1 model) can interpret and learn from effectively.

Deep Learning Model Training for Cross-Species Patterns

Leveraging deep learning architectures (like Evo-1), we train the model using the preprocessed data. The training process focuses on capturing and learning the genomic and expression patterns that enable successful cross-species gene function—specifically, how sequences from Rhodococcus erythropolis can be adapted to express efficiently in Streptomyces lividans. Model parameters are optimized to enhance the accuracy of predicting sequence compatibility across these species.

Cross-Species Sequence Optimization and Prediction

We develop a predictive module that accepts input features related to the target gene (including its original sequence, characteristics of the host Streptomyces lividans, etc.). Using the trained deep learning model, this module generates optimized gene sequences tailored for cross-species expression. It then outputs these optimized sequences along with corresponding confidence metrics to indicate how well the new sequences are expected to function in the target host.

Club Exhibition

We actively participated in a school-organized club exhibition to introduce our club’s mission and activities to a wider audience. We also designed posters to promote the iGEM competition and our project. As mentioned earlier, we hope to inspire more people to take an interest in science. You can see pictures at club page.

Speech

With the school’s permission, we organized a small-scale presentation. During the talk, we gave a detailed introduction to the iGEM competition and explained the principles of our project, such as how to design plasmid sequences and achieve conjugative transfer. We also emphasized the harm of mulch film residues to the soil and discussed how to balance economic benefits with environmental protection.

Offline soil condition check

We chose Henan Province to do investigation (Mianchi country, Sanmenxia), and based on our observation, the surface soil moisture was generally moderate to low, with occasional short-term drought.

Residual polyethene mulch fragments were widely present in the topsoil, varying among plots and often affecting water infiltration and root growth. Frequent mulch use but poor recovery appeared to contribute to unstable soil moisture conditions.

Data Analysis

This online survey, based on about 100 farmer questionnaires from Sanmenxia, Zhumadian, Zhoukou, and Shangqiu in Henan Province, focused on three main aspects: mulch film usage habits, awareness of residual film impacts, and willingness to adopt eco-friendly alternatives. The results basically indicate that plastic mulch was widely used, with conventional polyethylene films still dominating. Residual film management is generally rough, though many farmers have basic awareness of its hazards.