This section deals with more specific areas of my Ph.D. and some of the problems and difficulties that I encountered, which could have been prevented or been easier to deal with had I known about other peoples’ experiences. Having thought that I am quite an organised person, it was a bit of a shock the first time I repeated a set of experiments at the start of my Ph.D. Recording protocols in a form that I could follow, let alone that other people could understand, was a challenge that took some time to conquer. This is very important though – what is the point of carrying out an experiment if it can’t be repeated with enough confidence to prove the initial results?
From early on in my Ph.D. I kept detailed records of my experiments in my lab book with an index and filed data on my computer in a logical way so that any experiment could be identified and repeated with ease. Until you lose irreplaceable data, or sections of your thesis that have not been saved in more than one location, you can’t imagine how important it is to back up files.
Another lesson learned early on was that experiments without adequate controls are near to worthless. In a Ph.D. no-one can tell you what the result ‘should’ be, so you have to trust the data. For example, how can you tell if there has been any change in a readout following a certain treatment, if you do not have the data without treatment? How can you be certain about the kinetics of the response to treatment if only one time point is analysed? Time courses, along with titrations of treatment concentration, are key to investigating the progression of a response. Incorporating all of these factors into experiments makes them harder to do, but it increases the power of your results and will make you more confident when defending your data.
Experiment size can also affect your results. If too many conditions are attempted at once before you are confident of carrying out new methods, you increase the risk of making a mistake or taking too long and having potentially useless data. Equally, experiments should incorporate enough conditions for the data to be meaningful. A balance will come with experience, but getting that experience always involves making mistakes that can sometimes be as important as doing the experiment itself.
It is really important to check the reagents that you find in publications that you want to use in your experiments. A recent example from my experience was using a primer pair from a paper for a polymerase chain reaction (PCR). The paper did not say that the primer sequence was specific to one strain of mouse, and would be incompatible with genetic material from alternative strains. Following confusing results using these primers, I checked the sequences in an online program designed for searching the murine genome and this revealed that the sequences were unsuitable for my experiment. This was a lesson that I will not forget – publications may not give adequate information in the materials and methods, so you must check them yourself.
Carly Guyver.