Team:BM-AMU/Human Practices

Team:BM-AMU- 2019.igem.org

Human Practices



Background

Epithelial–mesenchymal transition (EMT) plays a crucial role in embryonic development, wound healing, and cancer metastasis. It is a progressive process with many heterogeneous metastable states in the middle, not only the epithelial state and mesenchymal state. These hybrid phenotypes are endowed with combined epithelial (e.g., cell–cell adhesion) and mesenchymal (e.g., motility) capabilities. It is clearly of interest to explore heterogeneous metastable states in EMT for understanding cell-fate determinations during embryonic development and tumorigenesis.

After identifying this valuable field, we are not trapped in the lab, but reach out to and learn from diverse stakeholders (Histology and Embryology Laboratory; Department of Pathology; Biotechnology company). Turning ideas into reality requires us to consider influence to communities. Therefore, we need to collect multiple opinions to ensure that our project is beneficial and responsible to the world. And furthermore, the feedback can somehow optimize our project, while we are trying to give out accessible solutions to what has troubled local communities through our designed two-way dialogues.

Inspiration
    • Output of inspirations:

      After the first input from diverse stakeholders (Histology and Embryology Laboratory; Department of Pathology; Biotechnology company), we came up with our 1.0-vision: Develop a method for accurately capturing cells in the intermediate states of EMT, so captured cells are not destroyed and can be used for further exploration. After the team brainstorming, we designed the project: EMT markers labeled with fluorescence as a precise monitor to measure the degree of transdifferentiation from E to M. Thus, we can accurately derive cells in different intermediate EMT states based on distinguished fluorescence expression.

  • Execution

    A solution to defects of traditional inducer.

    TGF-β is a classical controller for inducing EMT transformation, but we found that the process of EMT induced by TGF-β is unstable and discontinuous. In order to control the EMT process more accurately, we consulted Prof. Rui Jian again and got some insights.

    “In addition to inducing EMT progression, TGF-β has many other physiological effects, such as inflammation, tissue repair, and immune regulation. It acts like a master switch that controls many secondary switches below. ”- Prof. Rui Jian

    Feedback: we can find the end switch that manipulates the EMT in one direction and construct an orthogonalized system. After looking through the literature, we intend to use fundamental transcription factors as more precise controllers.

    Adjustment: We chose Snail1 and Twist1 transcription factors as controllers. Additionally, the Tet-on system and the Tamoxifen-Ert2 system were selected to control the expression level of transcription factors.

    A two-way dialogue

    unlocking the secrets of transcriptional landscape from cellular phenotype

    We have tried to design a software. When we input a fluorescent image with EMT markers labeled, it can help output the most probable sketch map of transcriptome. In order to popularize our software tool and get some advice for further promotion, we have organized an interview towards our potential users.

    Prof. Jiaxiang Xiong

    In order to evaluate the application value of this software, we met Prof. Jiaxiang Xiong, Teaching & Experiment Centre of Basic Medicine.

    Input: we’ve explained the mechanisms of image recognition and its applications in medical sciences, and convinced him that our software tool was practicable.

    Output: he was doubted about the users’ accessibility due to the difficulty of getting a fluorescent protein knock-in model.

    Adjustment: we are developing an update version to make immunofluorescent samples accessible.

    Dr. Ying Wan

    In order to evaluate the feasibility of software design, we communicate with Dr. Ying Wan, CTO of Vilang.

    Input: He agreed with us about that idea of predicting after our introductions.

    Output: Dr. Wan believed that our software is feasible, but he proposed that the credibility of the experimental results is not strong enough. Therefore, he suggested us to promote the algorithm to adapt high throughput samples in the future.

    Adjustment: Convolutional neural networks (CNN) models were developed here to expand the Supporting Vector Machine (SVM) models’ capability of handling high throughput data.