, , , ,

Question continued: I am a phd student in China.I just entered into the research group. But, my supervisor is very busy. I have to think out the project by myself. Now, i am very frustrated about research. The research papers are so many for me that i don’t know how to choose them. Finding a worthy project is like looking for a needle in a haystack for me. Besides, i am totally new to the research and i need to know how to go though the tough time.
So, i am here asking for your suggestions. You can take this opportunity to write a summary of your research life.
It can be the important points during choosing the project.
it can be how to read papers and arrange these papers.
It can be how to record your experiments in a better way.
It also can be how to get along with your supervisor, especially a busy and push one.

Tirumalai Kamala’s answer:

A 1st time researcher choosing a project

  • Initially choose to work on a method that has relevance to the Principal Investigator (PI)’s project.
  • Why method? Developing or optimizing a method, one learns the Scientific method.
  • It’s also practical. Unlike an open-ended scientific project with many beginnings and ends, a method is finite and most method projects could be finished within a year. In that time, not only would you become experienced in the techniques the lab uses, you may also have enough data for a method paper, if you’re lucky.
  • An ancillary benefit? Your PI and colleagues would benefit from your work. Win-win.

Here’s an example. A few years back, I had a terrifically talented and unusually motivated technician. She had a Masters in bioengineering and was new to biomedical research in general and to immunology in particular. One of the techniques we use a lot in my particular line of work is intracellular cytokine staining of T cells and flow cytometry to assess their cytokine responses. The method we were then using was cumbersome, typically 18 hour days or longer, starting very early in the morning until late at night.

With my inexperienced but extremely motivated technician, I mapped out an approach that broke down the process into a two-day protocol, i.e. a much more manageable 10-hour 1st day followed by a 5-hour 2nd day. One-day and two-day approaches needed to be compared side-by-side. Each change between the two protocols needed appropriate controls. Four months of focused work later, we had a working, validated (other teams generated similar data independently) two-day T cell intracellular flow cytometry staining protocol that worked so well everyone at our site now uses it.

How to read papers

  • A research paper’s most important parts: Experiment Design (what they did), Methods (how) and Results (what they found).
  • A research paper’s least important parts: Introduction and Discussion, i.e. the authors’ story around their data.

An excellent researcher learns to analyze and interpret the data for themselves, and does not depend on the authors’ spin. For this, basic interest in the subject and innate curiosity are necessary. Without them, the task is too tedious and laziness makes it easier to just follow the authors’ lead and take their story at face value.

  • Focus on Materials and Methods, Figures and Results. Pay less attention to the Introduction and Discussion. Initially very difficult but necessary.
  • Goal is to be able to reconstruct the essence of a paper simply by reading its Figures.
  • Try to read at least one paper per week with the goal of being able to reconstruct its essence from just its data, not from the authors’ Introduction and Discussion.
  • Start with papers published by your PI. After all, you are likely to work on a related topic. And you could easily clarify what you don’t understand by asking your PI and/or colleague to explain.

Once you get better at reading papers, you’ll find it easier to choose papers and organize them. Why?

  • Because once you find something in one paper that interests you, it’ll lead to a reference in that paper you find interesting. In turn, that’ll lead you to more interesting papers and so on.
  • It could even be something as simple as reading other papers by the author(s).
  • Read reviews to get an overview on a topic, and keep track of the work of the scientists who wrote them. They’re likely among the leaders in that field.
  • This is how you’ll learn to map and build your own database of scientific literature.

Ultimately, the best approach to learning the scientific method is through learning to effectively perform a critical analysis of data.

  • One of the best techniques for rigorously learning the scientific method is through a weekly journal club.
  • One person in a group presents a paper of their choice, and everyone joins together to critically examine the data, to find flaws in it or in the methods, and come up with alternative explanations for the results.
  • Many labs pay lip-service to journal clubs so few scientists are themselves good with the scientific method.
  • If you are committed to becoming an excellent scientist, try to participate actively in journal clubs in your lab/department/college/institute.
  • No journal clubs? No problem. Start one yourself with the help of other students in your lab/department/institute.

How to arrange papers
I use a common nomenclature for naming saved papers: year-journal abbreviation-1st author surname. An example? Page on nature.com which I saved as 2014-NatComm-Dutilh.

I maintain different topic folders. My topic sub-folders contain a topic’s sub-categories and so on. Here’s an example of my system for organizing papers, from microbiota, a topic of interest to me. Microbiota —> Microbiota sub-categories, one of which I named Virome —> Virome sub-categories, one of which I named Human Stool —> Human Stool papers on Virome.

Think of organizing scientific papers in your subject along these lines. Your job is to figure out a system with which to separate papers on a topic into its sub-categories. One way to look at the task? Like organizing a wardrobe. Also some papers may fit in more than one folder or sub-folder. In that case, I just save copies of that paper in each relevant folder.

How to record experiments

  • Record everything, everyday. Often, habits and laziness take over. That’s why it’s important to learn good habits early and learn them well.
  • In any lab, there are a set of standard assays and protocols. Then habit takes over and the scientist just records something like, ‘harvested cells, processed them using protocol #123′. Big mistake! No method is ever done the same way over and over.
  • Learn to record as you go.
  • You just need to develop habits that make it easy for you to record as you go. That means dividing your lab work into modules/chunks/units.
  • It’s easier to record data in real-time with electronic tablets in the lab rather than in physical lab notebooks. One advantage is that the entire experiment is electronically recorded in real-time and can be uploaded as a complete record into an electronic notebook, if your lab uses them.
  • As you’ll use the same methods over and over again, you’ll come to understand how to efficiently separate out different parts of the work. For example, steps one can do ahead of time and those only do-able at a specific moment on the day of the experiment. From this, develop excel templates for each protocol you use.
  • Cook on a routine basis? Think of lessons learned. Many cooking lessons are directly applicable to experimental lab work.
  • In turn, this’ll teach you to organize and divide the data recording into excel templates with pre-set formulas that are optimal for your work. This helps record data in real-time in a time-saving manner.
  • I’ve often noticed that people leave the lab as soon as they finish setting up their experiment. Big mistake! I train people who work with me to take an extra half hour or so to finish recording everything about their days’ work electronically while it’s still fresh in their minds. No coming back the next day to do this. Many details will be lost to memory.
  • What do I mean by record everything, everyday? An everyday example from my work life is a cell staining experiment followed by flow cytometry.
  • I teach colleagues to record every step of the protocol. It’s an electronic template. They just have to record real-time differences/changes from the norm.
  • With antibodies for example, I teach to record every item, including reagent name, antibody clone name, flurochrome conjugate, vendor name, catalog #, lot #, expiration date, concentration, amount used.
  • Using excel spreadsheets, one can make templates for recording this information and much of it can also be copied from one experiment to another with only minor details to be changed on the day of a given experiment.
  • Setting up detailed and specific templates to record experiments takes time and effort initially but once set up, they are very easy to use and maintain, and data gets recorded more rigorously and with more granularity/detail.

In my opinion, every lab should have electronic notebooks where data is recorded in real-time, date and time stamped, and uploaded to a central server, with little or no scope for post-modifications, only additions. There is far too much scope for fraud and misconduct in science these days, and we need widespread use of such electronic systems to pre-empt or minimize their scope.

Getting along with the supervisor
I don’t know if I can give much useful advice about this since cultural differences are likely to be considerable. With a busy and pushy supervisor, try to pay attention to patterns. For example, does the supervisor have meetings at specific times of day and/or week? If yes, capitalize on that pattern and put yourself on that meeting calendar, i.e. make a habit of meeting with your supervisor regularly. Go in with a bulleted list of action items. Update on your work? Issues getting research materials? Issues getting access to a paper? Point is to make a habit of regular meetings with your supervisor. Use those meetings to make the supervisor familiar and comfortable with you. In those meetings, point is to start with information your supervisor likely seeks or finds useful such as status reports about your work. Then ask for information or support you seek or need. Overall goal is to build a rapport of some sort with your supervisor. Weekly meetings are more beneficial than not.