16S metagenomics data normalization

Question

[junaidkori]

I have 16S amplicon metagenomics data i want to normalize my reads because number of reads varies between different samples. Can any body help me.

Answer 1

[robert.kwapich]

If you use QIIME then you could summarize your biom table, and choose appropriate depth through '-e' parameter in core diversity analyses. This accounts for differences in depth between samples. To choose this value you could simply create a rarefaction plot again in QIIME.

Just read on:

http://nbviewer.jupyter.org/github/biocore/qiime/blob/1.9.1/examples/ipynb/illumina_overview_tutorial.ipynb

and

http://qiime.org/scripts/make_rarefaction_plots.html

I usually sumamrize table and take the "sensible" lowest value. I.e. if the values are more or less for ex. 1000, and there is one around 150, I choose 1000, therefore excluding a particularly shallow sample and ensuring further normalization.

There may be other methods though.

[fanli.gcb]

Other methods include not rarefying: http://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1003531