HackMD
Virtual Meet-up: Bioinformatics Hub of Kenya initiative
2020-06-19
1. What is the advice to an individual interested in bioinformatics who has basics only? The key basic skills required for someone with no prior knowledge on bioinformatics?
- key skills:
- search, filter, extract, cross-reference data from large databases - make use of data & knowledge that's already out there!
- sequence alignment - core concept in evolution/phylogeny, functional genomics, genome assembly, differential expression analyses, transcriptomics, metagenomics, etc etc etc
- parsing data - reading data from many different (often messy!) file formats
- organisation - keep track of what/where your data is, which analyses you've run, with what parameters/settings, etc
- start with web-based tools e.g. EMBL-EBI/NCBI resources
- EBI Train Online has a huge amount of freely-accessible content introducing the fundamental concepts & guiding users on getting started
- once you begin working with larger amounts of data, you'll probably need to learn some command line computing (avoid long waits/costs of uploading data & downloading results)
- many great resources to learn the basics, e.g. Software Carpentry Shell & Data Carpentry Genomics
- The Galaxy platform provides a fantastic GUI alternative for those unfamiliar with command line computing
- Learn Galaxy and Galaxy Training Network have many excellent tutorials to learn the platform and bioinformatics simultaneously
- Galaxy can be installed locally to avoid upload/download of data over the Internet, but requires access to available server and some knowledge of server administration
- some understanding of statistics is also necessary
- Bernd Klaus' teaching material is a good place to start (if you know R) & Modern Statistics for Modern Biology by Huber & Holmes is a more comprehensive, but less accessible, guide to modern methods
- other good, free, online resources I know of for learning bioinformatics:
- H3ABioNet Resources:
- Simon Cockell's Lockdown Learning Bioinformatics-along videos
- Applied Computational Genomics from Aaron Quinlan's Lab at University of Utah
- for more, see "Teaching" section of http://quinlanlab.org/
- more here
- Finally: know that, if you're spending a lot of time searching the internet for help/answers, you're not alone!
- (search first, to see if your question was already asked by someone else!)
- http://seqanswers.com/
- http://www.biostars.org/
2. Is there a specific programming language preferred in the bioinformatics field?
- Python & R are equally popular and great places to start - free, open source, easy to install, huge online community, many resources to help you learn
- Choose whichever language your friends/colleagues are already using - I suspect this is the single biggest predictor of success
- Otherwise: Python is good for image analysis (so is ImageJ/Fiji, which provide a graphical interface), and more broadly applicable/useful outside bioinformatics, R has more cutting-edge statistical methods because of Bioconductor
- if using/learning Python: check out Biopython
3. How to build strong skills on a given programming language for data analysis and visualization.
- study other people's code - how do they do what they do?
- Python: learn numpy; pandas; matplotlib
- Use JupyterLab or Jupyter Notebook
- R: learn Tidyverse (dplyr; readr; tidyr; purrr; ggplot2; etc)
- Use Rstudio; work in RMarkdown
- Make it Open & Reproducible
- data analysis:
- Jake Vanderplas's Data Science Handbook (Python)
- Hadley Wickham's R for Data Science
- Wes McKinney's Python for Data Analysis
- sadly not free: eBook PDF (no DRM) costs €~34
- Rosalind programming challenges will help you to simultaneously develop programming skills and insight into bioinformatic algorithms & approaches
- for data viz: use an interactive environment like Jupyter or RStudio - makes iterating over/exploring new visualisations much more fun.
