Tweeting a lot about bad microbiome analyses, and got a few DMs from newbies asking about how to good analyses. It's hard to give a complete answer, but I'll share some tips below. Comment with your top tips too!
In my opinion, the *biggest* problems are not unique to microbiome. They are: lack of clear research question, data collected does not answer research question, sample size under-powered, multiple testing without proper FDR control
As for the problems that are more unique to microbiome. They are: "How do we normalize the data?" "How do we deal with data sparsity?" "How do we deal with high-dimensionality (more features than samples)?"
If you're new to the field, I think the best piece of advice is to say that it is a new technology so best practices don't really exist / are constantly evolving. I have my own pipeline, which I like best -- but there are many valid analyses.
As an analyst, you have the responsibility to THINK ABOUT WHAT YOU'RE DOING. Read benchmark papers, workflow papers, editorials, etc. Engage with the bioinformatics literature, and engage with bioinformaticians too.
When I read a paper, I want to know your justification for why each analysis was done. Tell the readers exactly what you did, but also tell them why you did it. Reference benchmark papers, workflow papers, guidelines, etc.
Finally acknowledge the limitations. Think about how changes in study design, sample collection, pre-processing, analysis, etc. could change the results? Interpret with caution, and don't be afraid to publish a null result.

• • •

Missing some Tweet in this thread? You can try to force a refresh

Keep Current with Thom Quinn

Thom Quinn Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!


Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @tpq__

2 Mar 20
Once again @badlytrainedtec and I have teamed up give you all a new method and R package for compositional data analysis.

How does it work? See below (1/N)…
@badlytrainedtec CLR and ILR are both popular for compositional data analysis. These are really helpful, but there are limitations. They both use a log-ratio transform of geometric means, and so (a) fail when you have zeros and (b) are not necessarily as interpretable as you think (2/N)
@badlytrainedtec Consider the balance (A & B) vs. C. If B > C, you would expect that the mean of (A&B) is always larger than C. This is true for the arithmetic mean, but it isn't true for the geometric mean! For the geometric mean, (A&B) can be much smaller than C if A is rare! (3/N)
Read 11 tweets

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!

Follow Us on Twitter!