We aim to teach basic data skills that give you freedom.
Running bioinformatics software isn’t all that difficult, doesn’t take much skill, and it doesn’t embody any of the significant challenges of bioinformatics.…These data skills give you freedom…
I believe these two qualities — reproducibility and robustness.
So what is a reproducible bioinformatics project? At the very least, it’s sharing your project’s code and data.
In wet lab biology, when experiments fail, it can be very apparent, but this is not always true in computing. Electrophoresis gels that look like Rorschach blots rather than tidy bands clearly indicate something went wrong. Unfortunately, without prior expectations, it can be quite difficult to distinguish good results from bad results.
The easy way to ensure everything is working properly is to adopt a cautious attitude , and check everything between computational steps.
You will almost certainly have to rerun an analysis more than once.
Write Code for Humans, Write Data for Computers
Use Existing Libraries Whenever Possible
Treat Data as Read-Only
Document Everything (-- Too geeky?) Just as a well-organized laboratory makes a scientist’s life easier, a well-organized and well-documented project makes a bioinformatician’s life easier.
-- <<Bioinformatics Data Skills>>
Yumin Zhu1, Gang Xu1, Xiaocheng Xi, Xupeng Chen, Zhuoer Dong, Xi Hu, Jingyi Cao, Siqi Wang and Zhi J. Lu*
1Contributed Equally *Corresponding to Zhi J. Lu
Lu Lab 鲁 志 实验室
School of Life Sciences, Tsinghua University, Beijing, China
e-mai: lulab1 AT tsinghua.edu.cn
Copyright © 2019 Lu Lab