Team:USTC-Software/Human Practices
Human Practices
Conference: iGEM introduction
At the beginning of our project, the core members of USTC-Software conducted a meeting of iGEM introduction. During the meeting, the team leader Shuyuan Shen introduced the targets and requirements of the iGEM competition. Then the frontend and backend leader, Xuanchao Peng, Haoyuan Wang and Longbang Liu, introduced the skills needed and the reference material to learn. Given the large scale of our team, tasks need to be scheduled and distributed properly to ensure the efficiency of the development of our project.
Discussion with bioinformatic researcher
In order to decide on the topic of our project, we reviewed the projects of the outstanding iGEM teams in previous years, and we found that the first thing we should do is to visit bio informatic researcher. So we contacted Junxi Yi from the Biocomp Lab in our school and got a better understanding of synthetic biology.
During the conversation, we learned that the traditional synthetic biology suffers from repetitive trail and errors, and they need predictors to help them determine some features of proteins in advance. So we suppose that machine learning may help with the prediction process, but we are in lack of academic achievement of machine learning on biology information.
During the conversation, we learned that the traditional synthetic biology suffers from repetitive trail and errors, and they need predictors to help them determine some features of proteins in advance. So we suppose that machine learning may help with the prediction process, but we are in lack of academic achievement of machine learning on biology information.
With the help of Junxi Yi, we contacted Professor Liu Haiyan, Principal Investigator of Biocamp, and arranged a meeting with him.
Meeting with Professor Liu
After the preliminary information collection, we met with professor Haiyan Liu in School of Life Sciences, USTC. Professor Liu introduced their research on computational tools for biomolecules and the current progress of their development. When mentioned machine learning, he offered us possible directions of our work and the possible issues we may be trapped during our development, and then gave us several examples on the machine learning model of other research group.
During the conversation, we realized that it may be not a wise choice for us to implement another model to predict protein properties. First, the difficulty of getting reliable databases is huge, and we are not the experts at model training. Second, the training process is some kinds of brute force, precise models often come with those clusters of computation nodes. However, we found that most of the models does not provide a pretty, modern frontend, and what's more, they are not integrated enough to allow researchers to experience at one place. Then the imagination of integrated site for those machine learning came up in our mind. Finally, after the discussion with professor Liu, we set the direction of software developing to the integrated web application.
During the conversation, we realized that it may be not a wise choice for us to implement another model to predict protein properties. First, the difficulty of getting reliable databases is huge, and we are not the experts at model training. Second, the training process is some kinds of brute force, precise models often come with those clusters of computation nodes. However, we found that most of the models does not provide a pretty, modern frontend, and what's more, they are not integrated enough to allow researchers to experience at one place. Then the imagination of integrated site for those machine learning came up in our mind. Finally, after the discussion with professor Liu, we set the direction of software developing to the integrated web application.
CCiC Online and Collaboration with SJTU-Software
At the end of August, we attended the Conference of China iGEMer Community (CCiC), the largest synthetic biology conference in China. In the online meeting, we introduced our project CAT, the Convenient and Accurate Tool, to the other teams in this conference, which was highly evaluated by many experts and researchers in the field of synthetic biology. During the conference, we listened to reports and advices from each team to further improve and enrich our CAT.
Furthermore, we successfully form a partnership with the SJTU-Software, the software team devoted themselves to training of machine learning model. After a deep conversation, we exchanged our ideas on machine learning service and they will provide an API for our website.
Hand Out of Brochures
After CAT was basically completed, we designed brochures about our project and related knowledge. By comparing the two methods of protein prediction - traditional biological methods and machine learning methods, the brochure provides a comprehensive overview of how CAT works and the merits and demerits of machine learning on protein prediction. By distributing the brochures in science and technology museums, middle schools, high schools, and in USTC campus, we spread the basic knowledge of protein prediction and machine learning to students at all levels as well as many adults, and received many positive feedbacks.
"This brochure showed me a lot of novel knowledge about synthetic biology and the iGEM competition, and I hope to do in-depth studies about biology and machine learning in the future."
-- From A high school student
Contact Us
Mail:
USTC_Software2021@163.com
University of Science and Technology of China, No.96, JinZhai Road Baohe District, Hefei, Anhui, 230026, P.R.China
