Career, Science

From Postdoc to PI — Ten Simple Rules for Applying (part 1)

Starting your own group is one of the most important steps in your scientific career — and one of the hardest.

Being invited to a Career Development Workshop at ISMB 2012 made me write down some of the advice that I had got when I was on the jobmarket a few years ago (and even put some of it on slides).

In a diverse and interdisciplinary field like computational biology it is very quite hard to come up with general rules that fit everyone. This is why I went down the self-indulgent route and revisited the CV and research statement I had prepared 4 years ago. (You’ll find a copy in the slides.) Some things are Ok, some things I would improve now — you will see, I’ll comment on this later. Let’s start with the basics:

1. Do a postdoc!

In some (maybe a bit more theoretical) fields like CS or Stats it is not unusual to go from PhD directly to a faculty position leapfrogging the postdoc. And there are several successful examples of this in Comp Bio too. So, is it worth doing a postdoc? Or wouldn’t it be much smarter to move straight from PhD to an independent position?

The answer is simple: Do a postdoc!

And maybe even a short second one if it can be used as an effective jumping board.

What you gain in a postdoc is scientific maturity. A postdoc is the best opportunity to extend expertise or maybe switch fields. Don’t stay in your PhD lab, you have already learned all the tricks there. Try a new city, country or even continent. You will learn new research styles and leadership styles, extending the toolbox you already have with essential skills for your own independent career.

What you need to look out for in a postdoc is independence (and you can help eg. by bringing your own money in form of stipends and fellowships) and support for your own ideas. Don’t be the data monkey. Don’t be one of 20 postdocs, who all compete with each other. No matter how famous your advisor is. Your environment matters, no matter how talented you are and how much you work. Without independence and support you will not be prepared for your own career.

Before you start a postdoc, check which percentage of previous postdocs in that lab have become PIs and are visible in the field. If an advisor has produced significant ‘scientific offspring’, that’s an indicator for a nurturing and supportive environment – just what you want.

2. Be concise! The curriculum vitae

Ok, now let’s look at the boring technicalities of job search. The obvious thing to start with is your CV. Here are some general comments:

  • Put your name and address on the top. Make your name stand out. Don’t write ‘Curriculum Vitae’ in bold – the people reading your CV know a CV is a CV without you telling them it’s a CV.
  • List your positions including the names of your advisors.
  • List your degrees. If you come from outside the anglo-american system you might think of translating your degrees into ‘PhD’ or ‘M.A.’ or whatever. Personally, I’m never sure what the right translations are. I list my degrees as I got them in Germany and as they read on my certificates.
  • List your awards and explain -very concisely- what each one means and how competitive it was.
  • Fill the rest of the page with the teaching and reviewing you have done.
  • After that on the second page list your publications. More is better. You can list submitted manuscripts. I’ve found myself also listing ‘papers in preparation’, which I now think is a bit cheeky. Especially since some of the things I listed never made it to print.
  • Keep your peer-reviewed papers separate from reviews and book chapters. Use a bold font for your name to make it stand out.
  • You can also list talks you have given. I spent 1.5 pages on that. Looking back I would select the highlights and cut this part to less than half a page (reducing the total length of my CV to four pages instead of five).
  • The last page is listing my letter writers and references. I chose my postdoc advisor, main collaborator during my postdoc, my two PhD supervisors and a professor I had visited during my PhD. Make sure these people actually know you. A letter is no good if it starts with the words “I have met Florian once for about 5 minutes …”, no matter how famous the letter-writer is.
  • In general you need a clear structure with clear headlines. Avoid white-space and condense the layout. Keep it at 4 pages or below. If you have too many talks/actitivities/etc select the 5 most important ones. As almost always: Quality beats quantity.

Update: Nature just had an article on CVs, which made me realize (i) that I have actually described a résumé and (ii) how important cultural difference are.

In the United States and Canada, a CV is comprehensive, whereas a résumé is concise. (…) It is important for international researchers seeking US positions to note that résumés should not include personal information or a personal photograph.

Indeed, when I sent my US-résumé (I’ve never needed a comprehensive CV anywhere) to a German university, they asked for personal information like date of birth, marital status and a picture, which I sent in on a separate page.

3. Be unique! The research statement

Your CV tells the hiring committee that you are a super-smart and hard-working person. Now is the time to tell them what you actually want to do once you have started your group. Start the research statement with a short intro into the field you are working in and the topics you are interested in. Four years ago I did it like this:

“My research is concerned with developing statistical and mathematical models of complex biological systems and analyzing large-scale molecular data. My research interests range from the analysis of microarray data in clinical settings to inference of cellular networks from high-throughput gene perturbation screens and integration of heterogeneous data sources using machine learning techniques and probabilistic graphical models.”

What rubbish! I must have tried real hard to put all the buzzwords in I knew. It sounds like I had no plan at all and was happy to do everything and anything. Now compare this with my current ‘mission statement’ that I quote from my webpage:

“The Markowetz lab for Quantitative Biology develops algorithms and statistics to leverage complex and heterogeneous data sources for biomedical research. Our main research question is: How do perturbations to cellular mechanisms shape phenotypes?”

Still not very specific, you say? I agree – but (a) it’s much shorter and (b) there is a research question. It’s better to state a question you are interested in than listing all the tools you know.

After this general introduction to the question you work on, your research statement needs:

  1. A summary of contributions during your PhD and postdoc. List of them as statements (“I pioneered approach X”) and then add a short and concise description to each item. Be quantitative and if possible mention the number of citations / downloads / site visits you received.
  2. Most importantly: your plans for the future. Make sure to conclude concrete short/mid-term projects and more long-term projects showing your vision of where the field is going. In job interviews I was asked several times ‘What is the title of your first grant?’ and I always found it helpful to point to one of the projects I had listed in my research statement as a way of showing that I had thought about such things.

To stand out it is important that you describe where your niche is (which includes, importantly, how you plan to distinguish yourself from your PhD/postdoc advisors). For the top jobs it’s not enough to just do what everybody else is doing – they want to see that you have the potential to be a leader in the field. And the best way to do that is to show your individual view and ideas.

… and 7 more points to come

I just spent more than 1000 words on the first three points, I’m a bit exhausted now. The other seven rules will follow soon.


Update: here is the follow-up post.

Image source:

7 thoughts on “From Postdoc to PI — Ten Simple Rules for Applying (part 1)

  1. I would be interested in hearing more on point 3, i.e. what was the process leading up to you defining your biological area of research?
    I have a physics background and learnt statistics on the job from an awesome PhD co-supervisor so I feel like I have the best of both worlds. and having spent roughly 6 years being outnumbered 100 to 1 by biologists, I am getting to stage where I can ask biological questions that many biologists aren’t able to, due to the differing educational backgrounds. However I will never be able to totally independently ask biological questions, it will always remain collaborative. But there are some problems I face sometimes:
    -how to move out of the consultation-in-return-for-acknowledgement loop and into genuine collaboration and experimental design;
    -how to effectively communicate the fact that everyday 3-d intuition fails in ultra-high dimension and therefore you need someone who is trained for this;
    -the success of R/BioC is a double-edged sword, why give your hard-earned data away for someone else to analyse when you can install the package and do it yourself? (the answer is: I read about e.g. covariance estimation for fun and can hopefully spot issues/pitfalls/better methods better than someone untrained).
    Anyone else have similar experience?


    1. Hi Kristen, thanks for your comment!

      I think you are right, I also think that “I will never be able to totally independently ask biological questions” – at least not the really interesting ones, but biology is nowadays such an interdisciplinary field that having a collaborative career is easy (and often fun).

      The trick is to start leading projects (instead of ‘just’ being the data analyzer). To get there, start proposing potential projects. If a project was your idea, you will find it much easier to get recognized when it comes to negotiating positions on papers for example. If you propose the project, you will also quite naturally be involved from the very beginning.

      And regarding you comment on BioC … I never had that problem. I’m actually quite happy if my collaboration partners give it a try first and maybe do some of the basic analyses themselves (because then I don’t waste my time on this) – if there are interesting problems in their data they will see their limitations pretty soon and be even more happy when you help them.


You gotta talk to me!

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s