Learning "Bioinformatics" in 2024

This is an update from my previous article Learning Bioinformatics in [2023], you should consider reading it, if you haven't.

I struggle with the term bioinformatics because it usually means different things to different people. So I'm careful when I use it and try to use it in context.

As an enthusiast, I am also interested in how learning is changing as technology advances. Writing is a good way for me to document the evolution of bioinformatics learning.

Everything I wrote in the previous article still stands. The foundations cannot be replaced.

I started working in a different area of biology, infectious disease and I'm still amazed about the amount of opportunities available for computational work.

The workflow of the same computational work 5 years ago is different from the one today. AI is changing how we approach work, how we learn, how we teach. We need to take advantage of the possibilities.

How can we leverage technological advancement to enhance computational biology learning?

Usage of AI/Chat bot to enhance learning.

Previous ways of writing code, was to learn the fundamentals and debug using google or stackoverflow. Now large language models like ChatGPT, Copilot are ubiquitous. It is inefficient to spend hours trying to solve mundane task without the use of these chatbots.

This frees up space for creative work. you have to use it and adapt.

Word of caution, these models make mistakes and sometimes perform poorly on creative or novel work. There are quite good for brainstorming and initial ideas generations.

Use a workflow manager

Starting out, I remember writing the same command individually and saving my results. Then I learnt about functions, and then packages. Even beyond this, I still forget, sometimes what version of the packages I used, or what parameter. Every now and then I'm lucky enough to have those information.

It was bad record keeping skills. Workflow managers like snakemake, nextflow help in reproducibility and automation. Now, I can easily access the history of previous commands i ran, I also don't have to run individual commands any more.

Project Organization and Documentation

Workflow management relates directly project organization.

A good project organization, is the one you can refer to in 10, 20 years time.

It is a foundational practice that we "computational people" ignore and I'll always take the opportunity to resound it again.

Keep it basic, a digital lab notebook, where proper documentation of test or experiments are recorded. If you do the same for wet lab task, there is no reason why computational experiments documentation should be different.

Worry less about what computer (Machine) I use

I spent to much time than I should have about what laptop to get for my work.

Now i connect to a high performance computing cluster for heavy computational task. With the amount of compute power available and cloud computing advancement. The Computer I use is one less problem.

Even storage, the compute power and storage available on the cloud is enormous.

That'll be all for now.

Keep Learning!