I am a Postdoctoral research associate at the Dementia Research Institute at Cardiff University. My group (Taylor) is interested in using single-cell omics to gain an in-depth understanding of microglial dysfunction at different phases of Alzheimer’s disease progression.

As a wet lab scientist turned Bioinformatician, I didn’t pick Bioinformatics up easily, despite my interest. There was a steep learning curve associated with transitioning into Bioinformatics, and it’s still something that I continuously strive to become better at. 

There isn’t a one-size-fits-all workflow, and with studies increasingly involving single-cell omics, multi-omics, and spatial transcriptomics it becomes crucial to understand single-cell best practices and stay on top of new developments. A big part of Bioinformatics is knowing your data, letting biology inform your analysis, ensuring adequate cut-offs and that the assumptions of the methods you employed in your analysis are met. The choices you make, for instance in how you normalize and integrate your data, can impact your downstream analysis and you need to be aware of that. 

In my opinion, reproducible research depends heavily on Bioinformatics data skills, version control and knowledge of statistics. It’s important to keep in mind that silent errors can arise quite easily so you need to adopt robust research practices that will make your analysis resilient against them. Reproducibility and robustness go hand in hand, and you can’t have one without the other.

Whether you’re just starting to venture into the world of Bioinformatics or are at a more advanced level, attending short courses focusing on a particular type of analysis is a great way to understand the different approaches and methods available and their limitations, as well as get some practical experience guided by experts in the field. In my opinion, being able to discuss your analysis with more experienced Bioinformaticians and learning from their experience and expertise is one of the fastest ways to improve your own skills.

Quite recently, I was fortunate enough to receive a BNA Credibility Bursary, which allowed me to attend an EMBL-EBI course on multi-omics data integration and visualisation. I highly recommend this course (you can access the courses’ resources here [1]) and its more advanced EMBL (Heidelberg) counterpart if you’re dealing with (single cell) multi-omics data. 

On top of continuously improving your Bioinformatics data skills, another great resource to increase the robustness and reproducibility of your research is the Turing Way Guide [2], which is an open-source, collaborative handbook for reproducible, and collaborative data science. I think this handbook should be shared as gospel for best practices in science. Go check it out, I promise it will make your science better!