As scientists, we spend decades perfecting our craft. Yet, most of that time is spent independently trying and failing in isolation, repeating the same mistakes that many have made before us. Here I will try to distill the "universal" lessons (learned through a lot of first hand failures) that new trainees will benefit from.
1. DO NOT DO EXPERIMENTS IF YOU ARE HUNGRY OR TIRED!
It's late, you're hungry, and way past bedtime but it is oh so tempting to just do a little bit more, get it over with, get that groundbreaking piece of data. Working with this mindset will likely result in experimental errors, which at best invalidates the experiment, wasting time and resources, and at worst, generates misleading data that leads you down unproductive paths for months or even years. Before you begin an experiment, ask your self - are you hungry? are you tired? If yes to either, address these problems first.
2. Document every experiment in a (electronic) lab notebook
Document EVERY detail especially deviations from previously established workflows. Often it is the details that matter in an experiment working or failing, and being able to compare every detail in retrospect is essential to pinpointing the key parameters that matter. Your memory can play a lot of tricks and is not to be trusted. An electronic lab notebook has the "find" function which is makes the lab book much more useful.
3.Make sure you are having fun!
Science is demanding and is not particularly rewarding in the societal metrics of money or prestige. What it can offer you, if you stick it out in the long term, is limitless fun, intellectual stimulation, and a better understanding of the world. Those who are successful at science are having fun. If you are not having fun, you will not be successful. Make sure you are having fun. If you are not having fun doing science you are doing it wrong.
4. Dust is your enemy!
This one is particular to microfluidics workflows. Anything that gets injected into a microfluidic device must be either a) filtered with a <5 micron filter or b) prepared in a dust-free covered workstation! Taking these precautions is definitely worth the trouble, as the consequences of a dust clog in the middle of a complex experiment can ruin your week. Dust clogs are not fun. You want to be having fun.
5. Beware contamination!
Separate high contaminant and low contaminant areas
This one is particular to single-cell workflows. Things that would typically be not a problem for regular experiments (such as a stray cell here or there, some stray DNA floating here or there) are catastrophic for single-cell workflows. I and many others have wasted months chasing down strange results resulting from contaminations. To avoid this unnecessary pain, separate all low contaminant (such as diluted cells, DNA, pre-PCR reagents, cell culturing reagents, buffers) from high contaminant (such as fully grown cell cultures, post amplification DNA) areas. This is typically done with a dedicated workstation for low-contamination preparation areas equipped with a UV light for decontamination. Avoid bringing anything that has been in a high contamination area into a low contamination area. Regularly decontaminate the low contamination area.