Top 10 Insights for New Researchers Starting Genetic Association Analyses
By: Juan Ramirez
Ever since the completion of the Human Genome Project, I’ve been excited about how discoveries from genetic studies could help doctors engage in Precision Medicine, where prevention and treatment efforts are customized based on an individual’s unique biology.
For new researchers, precision medicine can be not only exciting, but also overwhelming at times. Massive amounts of data are generated from multiple sources, spanning from genomics to transcriptomics, proteomics, and metabolomics. As the field moves quickly, it can also feel hard to keep-up on the latest methods.
As someone who has navigated these challenges, let me share my “top 10 insights” that have helped me – and will hopefully help any new researcher – start working on genetic association analyses.
1. Remember it’s okay to not know everything.
I cannot emphasize this enough. Part of the scientific process starts by recognizing what we don’t know. So, embrace your temporary ignorance and know it’s okay to ask for help. In no time, you will surprise yourself by how much you can learn.
2. Know your data well.
Knowing the methodologies of how your data were derived, which cohorts were included (in cases of meta analysis), and the unique characteristics of your sample (i.e., gender, age, etc), can save you from running more analyses than necessary. In fact, processing a large dataset very often takes more time than performing any analysis! Keep in mind that public data can have technical biases, which can arise from steps at library construction or from the genomic content itself due to various sample representations. So thoroughly read the supplementary sections of the papers that cite your data. These often contain lots of very helpful information.
3. Understand the statistical tests you employ.
Correlation does not imply causation. You heard this, right? It’s one of those phrases that stays with you. In genome-wide association analysis, your knowledge of statistics will be continuously tested. Figuring out how genes and phenotypes of interest are connected is key for adequate inferences. You cannot just rely on being a good programmer. Having a greater understanding of statistics will help you become a better researcher.
4. You can probably avoid recreating the wheel.
Luckily there are plenty of online collaboration and programming communities where people share their tools and work through problems together. In my experience, these are often great resources because you can connect with other analysts facing similar difficulties (e.g., processing some data in a specific manner, running a program, interpreting methods or results). People are welcoming and open to review your code and give good advice. Some of my favorite sites for bioinformatics discussions are: seqanswers and reddit. Github is increasingly being used as many developers post their programs, which can be easily downloaded with examples files.
5. Keep a lab notebook.
Since electronic documentation is the standard way of record keeping, you will need to formulate a way to keep track of changes to your code. Chances are that due to the various sources of data you will be using, you or any one of your colleagues will be coming back to repeat an analysis and recycle your code. Your future self will appreciate any comments/directions you left in the scripts, and any other additional notes left to clarify the name of a variable or why you used the data a certain way. Keeping detailed notes will save you from hours of detective work.
6. Discuss problems with your colleagues.
Scientific inquiry requires good communication, not just online but with the people around you. Most of the things I have learned have been through conversations with members of my lab or, other labs. When I first joined the Dunn Lab, Dr. Dunn encouraged me to get to know everyone by having “coffee dates” with my lab mates. My lab mates have proven so helpful in providing direction in those moments when I feel stuck.
7. Since you have to sell your ideas, you need to write A LOT.
Scientific research is not only about analyzing data. It’s also about communicating the results, and applying for grants to test new ideas. Writing effectively for scientific and non-scientific audiences is therefore essential. Although it may be challenging for some of us -- particularly if you are non-native speakers, one way to improve is by practicing. So get to writing. In time, you’ll be able to communicate complex ideas for all kinds of audiences. Happy writing!
8. Find and attend workshops and seminars of interest.
It’s important to stay up-to-date with current technologies and standards for research. Often, your university or institution will sponsor workshops, seminars, or trainings (announced in emails or flyers) where you can check out what’s going on in other labs or learn new skills to boost your work efficiency. Although it can some times feel hard to pull yourself away from work to attend these kinds of events, you should. These often provide opportunities to ask questions and get answers right away. It will also give you a sense of where technology is heading, so you stay ahead of the game by knowing what’s coming next and being ready for it.
9. Be curious and try online courses.
If you like teaching yourself new things, there are many platforms online providing courses related to a number of fields, including genomics and data science (e.g., Coursera, edX, and Datacamp). The great thing about these online courses is the structure of lectures: there are videos, readings, and environments for practicing coding. More interestingly, you will often see that these courses are taught by leading researchers of their fields. You can complete these at your own pace. And, if you’d like, you could also sign up for a certification module without breaking your wallet.
10. Take advantage of Happy Hours.
Take a moment with your peers to chat and relax. Happy hours are good opportunities to meet folks and create opportunities for collaboration. History has proven that inspiration comes in the least expected moments. Wouldn’t it be nice to have some of those eureka moments during the popular happy hours with a glass of beer or wine in hand ready to celebrate in good company? You worked hard for it, enjoy it.
Just like precision medicine, you will know what habits will exclusively work best for you. I hope these insights are helpful to you as you get started.