Microbiome data analysis: plans for curriculum / training

[bebatut]

We plan to expand our catalogs of tutorials for whole-genome microbiome data analysis.

Here are some ideas we will implement:

• Expand the introduction slide deck to add more overview of the different types of analyses
• Create a series of “small tutorials” that could be then used in a larger learning path.
  • Each tutorial could show 1 step in microbiome analysis
    1. Quality control and preprocessing (host data filtering, rRNA filtering, etc)
    2. "Core" steps (1 tutorial for each)
       • Assembly
       • Taxonomic binning
       • Taxonomic profiling
       • Functional annotation
       • Other?
    3. “Statistical” / downstream analysis (1 tutorial for each)
       • Diversity analysis
       • Combination of metagenomics and metatranscriptomics to get expression information
       • Multivariable association between phenotypes, environments, exposures, covariates and microbial meta’omic features / differential analysis
       • Antibiotic resistance detection
       • Other?
  • Content
  • Explain some of the theory behind the type of processing
  • Run 1 or 2 possible tools
  • Explain the outputs (including visualizations, etc)
  • Show metrics to evaluate the quality of the outputs
  • Point to other existing tools and (if existing) comparison of them using e.g. CAMI challenge results or other benchmarking
  • Focus less on the biological interpretation and more on explanation of the theory
  • Data
  • Same datasets for the tutorials in a and b
  • Different datasets (different purposes) for c (but explaining the steps taken to get the data)
• Create/Expand tutorials to illustrate a full analysis / examples
  • Content
  • Explain the input data and the original questions
  • Run the different steps with just short explanations and then linking to dedicated tutorials for the steps
  • Focus on the biological interpretation of the steps
  • Possible tutorials
  • Integrative meta-omics analysis - Metagenomics, Metatranscriptomics, Metaproteomics?
  • Metatranscriptomics (already available)
  • Cloud metagenomics / metatranscriptomics
  • Other???

Feel free to comment to add your thoughts, your wishes, and/or if you want to work with us on that!

[hexylena]

xref https://github.com/gallantries/organisation/issues/16

[hexylena]

I love the idea of smaller tutorials + learning paths to combine, in general I'm really enthusiastic about that sort of option, especially since it gives us options a bit like de.STAIR and "pick and choose which tool" to make your workflow and maybe compare the results a bit.

[millen2223]

I would be glad to work on that, too. The short version available in Metatranscriptomics is like a model version, right? I am still overthinking the plan points, but I find good.

[yenifergalarza]

I consider the large option is the best because it gives a complete perspective of the process, therefore the target group is the kind of people that usually have to read extensive papers. In another way for undergraduate students or massive targets, the first option could be the best.