handbook / Research Practices

Framing a Research Question

When it comes to identifying research problems, I take inspiration from Nobel laureate Jennifer Doudna:

“I think you can put scientists into two buckets. One is the type who dives very deeply into one topic for their whole career and they know it better than anybody else in the world. Then there’s the other bucket, where I would put myself, where it’s like you’re at a buffet table and you see an interesting thing here and do it for a while, and that connects you to another interesting thing and you take a bit of that.”

I view our group’s work largely through that second bucket. However, intellectual freedom always operates within realistic boundaries: available funding, formal project mandates (e.g., if you were hired for a grant with fixed milestones), and computational resources.

We also need to be thoughtful about domain overlap. Transitioning into an unfamiliar area, such as running biomolecular simulations with only a pure physics background, is entirely doable, but we will pursue it only when there is a compelling, strategic rationale to justify the steep learning curve.

When formulating your research question, use the following principles as your compass:

1. Ground Your Work in Why, What, and How

The “Why” (Purpose)

You should always be able to articulate why a project matters. Different motivations are equally valid:

Targeted mandates: Collaborative projects with predefined deliverables.
Open threads: Direct follow-ups addressing unanswered questions from prior work.
Skill exploration: Testing an emerging tool or technique to expand your technical skills.

None of these reasons is inherently superior to another, and your motivations will likely shift throughout your time here. What matters is that you have a rational justification that makes sense to you, to me, and to anyone you would stumble upon in your academic career.

The “What” (Novelty)

Look closely at what the literature already covers and identify what you can do new. In our (current) field, it is common to pick a material, compute standard benchmark properties, and publish. We actively avoid “run-of-the-mill” studies unless there is a strong reason, such as an unstudied phenomenon, an urgent technological application, or an onboarding exercise to master some tools.

Every project should clearly contribute to at least one of three areas:

Your growth: A rigorous chapter or core capability for your thesis/project work.
The field: A conceptual, physical, or algorithmic insight.
Our group: A reusable workflow, dataset, codebase, or method that can be useful for some members in future.

The “How” (Methodology)

This is where you have maximum agency. On the contrary to many resarch groups, we don’t want to stick to one set of tools, methods, etc. and would like to actively evolve our toolchain. Whether you use an existing packages, write in-house scripts, or adopt emerging tools, you have full freedom to choose your path, provided you can technically justify your methodology and validate your results.

2. Scope It Small First

Some projects are inherently ambitious and it is understandable. A good rule of thumb is 30% of the total research done in the group belong to that category. Rest we would like to have projects with targeted deliverabels within a maximum period of 1 year. This is because, from my own experience and others, it is very likely that you will loose interest on a problem if you keep working on it a long time (there are people and problems which are strong exceptions to this).

Particularly during your first 1–2 years in the group, we will prioritize well-scoped projects so you interest/expertise, publish early, and avoid research burnout.

3. How We Collaborate: My Role and Yours

I do not expect myself to automatically know more than you about your specific topic. As we pivot into new domains, you will quickly become the expert on the literature and other aspects of your domain. We can/should discuss (& discuss & discuss) and try to develop a thought one of us into a practial, acheiveable resrach task.

Our collaboration follows a natural arc over your time in the lab:

Early Stage                     Mid Stage                       Senior Stage
[ PI Guided ] -------------> [ Joint Co-Design ] -------------> [ Independent Lead ]
High input on design,         Discussions,                     You propose directions,
tools, and literature         Joint troubleshooting           PI advises and supports

In general, I would like everyone to come up with their own ideas as I think this is the best way to be confident, explorative and have fun along the way. But I would be happy to support in all the ways I can including, but not limited to regular discussions, manuscript writing and technical solutions.

← Literature Study