Starting a lab
My lab is 10 years old now, and quite a few of the people who have shared this journey with me are starting their own labs right now, or have done so in the last year or so and have foolishly been thinking I can give them some advice. I thought I would write down some of the experiences I have had. Most of it is probably obvious or wrong.
Finding your questions.
You probably had to tell your hiring committee or fellowship panel what question you are going to work on. If this is still the question you are working on 10 years later something has probably gone wrong. You need to find your question and it may take a long time. I am not sure I have found mine yet, but 10 years in, I am for the first time able to articulate a direction that is not just a different take on my post-doc advisor's questions. Looking back, I guess there are a few things to say about the journey.
Experiments take you towards your question gradually, and they do so through a growing sense of dissatisfaction in your own results combined with ideas that you see from other researchers (often working in very different fields from yours). Each time you do an experiment, you try something more ambitious. Each time you learn something but mostly you learn a new question. It sits there in your mind as a germ for months or years, until you notice something at a conference or in a paper that tells you how you might design an experiment to answer it. Colleagues will email you and say "let's get together and design a study". In my experience, no good study has been designed that way. All the good experiments come from ideas that have been bugging you and returning to you again and again, combined with sparks that give you insights into how to develop them.
So: Jump in and do experiments. Don't wait for perfection. Jump in with enthusiasm and gusto, but be critical of your own experiments and use your dissatisfaction wisely. When there is something you want to know, write it on your whiteboard so it stares you in the face every day. Go to talks and conferences outside your comfort zone until you find a way to answer it. If there is no way to answer your question with the techniques you know, find friends with different expertise. Don't be scared to approach people with your question. Everyone is looking for good questions, even people with awesome technical skills. Mostly though, jump in with gusto and don't be scared of experiments because of what other people think. The only thing that matters is whether it is taking you in a direction that you are interested in. Ten experiments later, you might have a question that is really yours.
If you ever become Buzsaki, Deisseroth, or Dolan you can have people that work for you. Until then, you work for them. You might think that paying someone $50k per year entitles you to tell people what to do, but they have put their futures in your hands, and your future is intertwined with theirs. This is particularly true when you start your lab. The people who join you have jumped over a massive river of doubt and taken the brave decision to work with you rather than a big cheese. You owe them. There will be many times when you will be thinking to yourself "I wish they would just fix it"- it is only human nature - but it won't work for anyone. Sit down and fix it with them, even if it takes weeks.
I think it's important to say that this happens to everyone that works with you. At some point, usually about two years into an experiment, everything will have gone completely wrong and this is no reflection on the post-doc or student. At this point, the worst thing you can do is get frustrated with them. It happens to the brilliant ones and the others just the same. At this point, it is your job to be completely confident that everything will be ok, and spend all your energy making sure it is.
If you do this enough times, one day they will turn up at your door with a completed draft and you will think "Wow – that one was for free. I made a scientist". That moment feels very good indeed.
Finding people to work with you
This might change as your lab gets bigger, but when you start out the only important question is whether you like spending time with your lab. People will differ about this but my take is that you want to start small. Because you don't have your question yet and because you want to give them all of your time and energy, you only want one or two people to start with. These people will be the majority of your working life. It will be tempting to hire people with particular technical skills or with a particularly strong track record. Of course these are all good things, but most important of all is that you can have a dynamic and interesting conversation that will spark into new directions and ideas. Don't hire the student from the awesome lab who you don't quite understand, or the technical guru who can do something you can't. Follow your instinct and hire the one you spark off; the one that challenges your thinking. That spark, that challenge, is the most valuable thing you can find when you are searching for your direction. When you have found your direction and you want something done, you can hire the others. If the spark isn't there in your interviews, don't worry (and maybe don't hire). It will be there in the next student intake, or at the next conference.
Finding funding to pay them.
Everyone will tell you that your job is to find funding. But, as everyone knows, finding funding for a new lab is really hard. The thing that nobody tells you is that spending money is also really hard. You won't believe how much money there is in a grant, and you won't believe how hard it is to recruit people to spend it on. If you get a 5 year grant, it might take you the first two years to recruit the person you want. If you get more grants, you will rapidly start spending your whole life trying to recruit people and manage them; writing references for their next jobs; making sure they have papers to write and students to manage so their CVs look strong. The easiest way to keep all this at bay for your formative years is not to get lots of grants. Both writing and managing grants is extremely boring and time-consuming. Don't do it for the sake of it. Remember that money is the input to science, not the output, so if you don't need the input, don't apply. There will be lots of opportunity when you are 5 or even 10 years in to get more grants and to expand your lab. At this stage, you will have the beginning of a question to work on, and you will have a track record to back up your applications.
Enough of my nonsense for now. I may update this when I think of other things.