Monday, September 4, 2017

How much time should a new PI spend at the bench?

Some time ago I saw Huda Zoghbi give a talk describing her career path and mentoring philosophy.
Huda is a giant in neurogenetics and neuroscience in general and she is one of the people who shaped the study of neurodevelopmental disorders. She is also well known to be a great mentor, in particular to women. One piece of advice that Huda gave to young investigators was to be at the bench all day every day, work closely with your lab and fill the lab with grad students. As a young PI you are the best postdoc you can get and you are the one who can continue to make the big discoveries. Grad students will take a while to train, but the time taken will pay off.

In theory, I would love to do that. But my question is "How?" How could I have responsibly hired 3-4 students in a very small program where I had 3 rotation students in 3 years? The start-up also does not cover the time necessary to graduate multiple students. With the fluctuations in funding, I could initially guarantee 2-3 years in salary for people and now I'm down to 1 year, so I have stopped taking rotation students altogether.  I have spoken to multiple friends in the same situation and some of them also have to teach one or two classes per semester giving them even less time for mentoring students.

For this reason, I have ended up running a postdoc and technician-heavy group, but as the tenure crunch is getting closer, I'm wondering if I should jump in to make things move faster. In talking to friends who just went through the process I realized I'm not an exception. I'm still the most advanced and technically skilled postdoc I can get. So my dilemma at this time is whether I need to get better organized with writing grants and papers, and carve out time for experiments. I was independent as a student and postdoc working on my own projects and ideas, so I tend to follow the same model of letting everyone work on their project and paper. Yet, I'm worried this is hurting more than helping and that I should get "all hands on deck" on a couple of papers and reroute resources to get things done.

What is your experience? How much time should a new PI spend at the bench? Does it make sense for me to step in to help with experiments or to start the riskiest projects and only give them off to trainees once the preliminary data is solid? Comments and advice would be greatly appreciated.


  1. 0. Get away from the bench. This isn't your job anymore. Set the conditions for productivity without you locked down as data monkey.

  2. Having started up my new lab exactly 1.5 yrs ago, I spent the first few months at the bench alongside my tech and postdoc (as probably most new PIs do). Then I handed it over to them as I got busy with grant-writing, finishing up manuscripts, etc. Just a few weeks ago, I had some free time so I jumped into the lab again and finished some smFISH experiments from start to finish to contribute to my postdoc's project. I did this because my postdoc and tech have their hands full. If I have more free time, I will jump in again to get things done and move the project(s) forward. If not, then I won't stress and will let those in the lab do their thing. This is something that really depends on the circumstance and there's no binary decision to be made.

    Every once in a while, I do go with the "all hands on deck" approach where everyone's efforts are corralled onto one project or experiment that needs to go into an upcoming manuscript submission/revision or grant applications with a strict deadline. Then everyone goes back to what they were doing.

    In terms of risky or not-so-risky experiments, I have worked hard (psychologically) to stay away and let the trainee learn and fail and keep learning (it is very tempting to just get it done yourself). I have seen tremendous improvement in their techniques and thought processes over time, as a result. Hence, my opinion is to give trainees the space and time that they need, unless of course you are facing a major time crunch and must do it yourself.

  3. I still help out on experiments about 10-15 hrs per week. It's lots of fun, great procrastination fodder, and let's me help students and postdocs get unstuck sometimes. Downside is I can also be a bit controlling and I'm always fighting the urge to micro-manage. I suspect my being in the lab adds a bit of pressure to some folks, especially the not-so-productive ones. Not sure if that's a good thing...

  4. When I started out, I spent about 10 hours per week in the lab. By my third year, I was spending next to no time in the lab. Between all my other responsibilities, and the fact that very little the high research skill work we do can be done in small chunks of time, it was no longer worth it. I analyze the data with my students (even now), but I don't spend much time in the lab itself.

  5. I am in the same boat as Prodigal Academic right now. I used to spend time in the lab and last summer worked 12-hr days to get a paper published, but right now between grants, teaching and other administrative stuff, I can barely get out of the office. Part of me misses it and another part realizes it's very hard to be productive when you only have chunks of time here and there...

  6. When I started, I was in the lab most of the day every day for the first two years. This was the most productive period my lab had. As my administrative/service/teaching has increased, I've seen a correlative drop in our data output the last couple of years. So I actually just set up a bench for myself again and I am forcing myself to find the time to do lab work and just physically be in the lab. Hopefully this gets us churning out data fast again.

    As for students, I've found that they are the best way to staff a new lab. First, they can more easily get their own funding than a postdoc via TA'ing, training grants, university fellowships, external fellowships, etc. I've taken all the good students who have rotated in my lab (3) without really worrying about where the money would come from. By applying for everything we could possibly apply for, it has actually cost me very little so far to have students in my lab. Students are also different from postdocs in that they're in tune with the happenings at the university and have a built in network of classmates to borrow reagents from, get advice, etc. I've also found that students want to be more involved in departmental seminars and activities than postdocs. And finally, the postdoc pool willing to work in a new, non-famous PI's lab is questionable at best. There are always some rare gems who could end up in your lab because of just the right set of circumstances, but these are incredibly unusual.

    1. "Students are awesome because they are cheap labor".

  7. I did a Western the other week. It was the first time I touched a pipet in two years and everybody was anxiously waiting for me to screw up (it was beautiful). I pretty much stopped being in the lab within the first six months. I don't even know how it happened, but I just became unreliable: I was always in meetings or teaching and there just never was a stretch of timeslots in which I could do anything useful anymore. Mostly things have worked out and since a lot of my projects were long-term things anyway, I didn't really feel too bad about the lack of immediate output. Now (4 years in) I'm starting to get worried though. But I think I will solve it by maybe being more on top of my people and becoming a bit stricter in my demands/expectations. My schedule is simply too full and scattered to make me a reliable presence at the bench. Also, I never really liked pipetting all that much, to be honest.

  8. It's my 2nd year as PI and I spend a lot of time in the lab (40h+ per week). There is absolutely no doubt that I am the most productive lab member at the bench. This will be true for most junior PIs - unless you are incredibly lucky, you will not be able to recruit the cream of the crop in the first few years. I am also fortunate to have very few responsibilities outside of benchwork and supervision of my team. Where I am, there is no teaching, grants are relatively easy to come by, so there is no need to submit 5 applications every year, and I am not well-known enough to be constantly swamped with paper and grant reviews. The benefits of staying at the bench in the first few years are really great:
    1. You can easily establish lab culture, both in terms of behavior, and in terms of technical details of experiments.
    2. You can set an example of high productivity that everyone else will be trying to emulate.
    3. You gain good will by being "in the trenches" with the team.
    4. You have an eye on how lab members interact with one another and can spot trouble early.
    5. You can more effectively troubleshoot with people - if you've done the technique and it works for you, then you know that the failure of one of the lab members' experiments is not due to equipment or reagent failure. This is very important with inexperienced lab members.
    6. You will find little ways of improving workflow that lab members might not come up with on their own - if you are annoyed by the little everyday snags there is a good chance that everyone else is, too.
    7. You can spearhead risky or technically difficult projects where it's really important to have an experienced pair of hands. Otherwise, some high-yield projects may end up getting dropped just because the person responsible couldn't handle the technical problems.
    8. If s#it REALLY needs to get done, it gets done (with your own hands).
    There are also drawbacks, but I think they are outweighed by benefits:
    1. You can be perceived as micromanaging and always watching lab members - some feel uneasy having the boss always around.
    2. It will to some extent limit the individual development of your lab members and their ability to troubleshoot on their own - if you are always there to help, it's easier to just ask you for help. This is a fine balance - where do you help out and where do you just have them fail and learn from mistakes.
    3. Obviously for people who are burdened with teaching, grant writing, service, and administration, staying at the bench will really limit the time they can spend developing new ideas, staying on top of the literature, and supervising lab members. If your time is really stretched thin, I agree that bench work may need to take a back seat.

    1. I think your comment sums things up perfectly. There are tons of benefits of working in the lab, but one must be careful about allowing lab members to grow. At the same time burden of teaching and grants and service can really kill you and make you feel like you don't do science any more...

  9. Hello, i am a very new PI (one phd student for 10months and 2postdocs for 5months) and I am already troubled by this too. I don't feel I can contribute much to help the discussion here due to my lack of experience. I would just like to say that I too struggle with the same thoughts. I am afraid too be to far from the lab,especially with the student but at same time I definitely do not want to micromanage, but the time runs too fast...however we (I) can always manage to think that maybe I it is just that I have to put more hours in and suck it up fir as long as necessary...
    Yet, we are just human...
    I appreciate reading your blog very much and today's post goes right to the gut of what I am thinking lately. I will be all over these comments too trying to learn a little.

  10. In my first few years I spend a good amount of time at the bench. 25+ hours per week. Just before tenure time I withdrew more and more. The standards in the lab, and particularly the "second crop" of grad students, suffered as a result. I'm now an associate prof and have set my self up in the lab again. I can only dedicate about 10 hours per week, but it makes a huge difference. I plan to keep this up as long as possible. IMO the benefits are immeasurable and somewhat intangible. Plus, I hate working on the computer and going from meeting to meeting non-stop for hours. The lab is a way to 'unplug' and clear my mind.

    The attitude that 'profs at my career stage should not be doing this sort of thing' really rubs me the wrong way. To each their own, but there are few things that annoy me more than colleagues pretending they are too 'important' or too busy to get their hands dirty. These are the same people who relish in telling you about all of the high-level crap they're involved in - from NIH study sections, to university-wide committees, to managing their start-up, etc. etc. I do all that stuff too, but don't waste time yapping about it. You'll still find me setting up reactions and teaching first year grad students how to properly purify proteins on any given day.

    1. Nice framing. It isn't about being too important or good for bench work. It is that this isn't your job anymore. The quicker noob PIs grasp this, the better for them and their research program.

  11. I am right around the time for promotion also and I still spend 20 hours/week at he bench. I am not the best "postdoc" in the lab anymore but I do it because I like it and it helps relieve some of the pressure in dealing with grant and paper writing -- I can get instant gratification from bench work much more often (however small the success). But I am not sure if this is the right approach.