Students in one of the graduate versions of this course will complete a computational biology research project. See the table for clarification.
The project is designed to give you practice in applying computational methods to contemporary problems in biology. Students design and carry out projects working in a group or by themselves. There is grading information below the steps of the project.
Steps of the Project
The due dates of the steps are listed on the calendar.
MIT students taking the course should email your background and interests to course staff. We will collate these replies and post them on a private area of the course web site to help with team formation.
Submit a list of your team members, a title, and a one paragraph summary of the research problem you plan to investigate. Groups can range in size from one student to five students. We suggest that you choose people to work with who have skills complementary to yours.
For a one-paragraph summary of your project, you might look at these sample applications, especially the first paragraph of specific aims pages. We encourage you to choose an area that is relevant to your current or planned research. If you do not have data available to use for this project, a suitable publicly available dataset may be found. Existing implementations of computational methods or custom programs may be used.
Each team submits a one-page set of NIH grant-style specific aims that describe your project. Your aims should correspond with goals you can reasonably achieve during the term. Your aims will be evaluated and returned with comments to give you initial feedback on your formulation of your problem area.
Each team submits a two-page (single spaced) research strategy describing the specific software that will be used or developed, the specific data that will be analyzed, any relevant statistical methods, the sequence of research activities and a brief comment on anticipated findings.
Each team submits a five-page (single spaced) written report that summarizes your findings. (Figures are included in the five-page limit, but references are not.) Each member of your group should author a clearly identified section of the report with their specific contribution. This report will be graded as described below.
Your research strategy and results section should be structured to include the following sections, with the bulk of the text in the Approach and Results section. (If you have multiple Specific Aims, then you may address Significance, Innovation and Approach for each Specific Aim individually, or may address Significance, Innovation and Approach for all of the Specific Aims collectively.)
- Explain the importance of the problem or critical barrier to progress in the field that the proposed project addresses.
- Explain how the proposed project will improve scientific knowledge, technical capability, and/or clinical practice in one or more broad fields.
- Describe how the concepts, methods, technologies, treatments, services, or preventative interventions that drive this field will be changed if the proposed aims are achieved.
- Explain how the work challenges and seeks to shift current research or clinical practice paradigms.
- Describe any novel theoretical concepts, approaches or methodologies, instrumentation or interventions developed or used, and any advantage over existing methodologies, instrumentation, or interventions.
- Explain any refinements, improvements, or new applications of theoretical concepts, approaches or methodologies, instrumentation, or interventions.
- Approach and Results
- Provide an introduction to the area, describing previous work.
- Descibe the overall strategy, methodology, and analyses to be used to accomplish the specific aims of the project. Include how the data was collected, analyzed, and interpreted.
- Describe the results of your project in terms of your aim(s).
The page limit is to ensure that the report focuses on the key aspects of the problem method. Section (C) 1 should allow a colleague who is not expert in the specific research area to understand the report. Nevertheless, it should take up no more than one page. Figures or tables may be helpful.
Formatting: Use 11 or 12 point font with at least one-inch margins, page numbers at the bottom, single spacing. All figures must have legends.
At the end of the term each team will present a short summary of their work. Groups will be allotted three minutes per person (e.g. three minutes for a group of one, 15 minutes for a group of five), followed by up to three minutes of questions per group. Overall, the group should include as many of these topics as possible:
- Research questions; specific aims
- Data sources, computational methods, and challenges you ran into
- Further questions
These presentations will be graded by the instructors and comments will be provided by other students.
All students will provide online feedback on the presentations.
Projects will be evaluated and graded based upon the following criteria. We will not be grading based upon the significance of your results, but rather on the quality of your thinking and execution. The success or failure of your chosen method in finding significant results should be thoughtfully examined in your presentation and written report.
Provide a clear and crisp definition of a biological problem that you have addressed with a computational method. Avoid generalities in specifying the problem. Specific questions to think about:
- Is your definition of the problem and the associated data clear and complete?
- Have you clearly articulated your assumptions?
- What are specific pros and cons of the problem formulation, the data?
- What controls have you suggested to avoid misleading results?
Provide a clear description of the computational method that you used. You should describe how the method works (the algorithm, any necessary preprocessing), which implementation details are important, and how one should interpret the results. It is especially important to try to understand how the method might fail in solving the problem. Specific questions to keep in mind:
- What is the computational criterion that the algorithm seeks to solve?
- What assumptions does the method rely on? When is it likely to fail?
- What controls could you use to ensure that the result is meaningful?