M. S. Binoj Kumar

The paper introduces a method to infer differential richness in 16S rRNA surveys by accounting for spurious taxa generated during sequencing, using an R package called Prokounter.

The paper addresses compositional bias in sparse sequencing count data by proposing an empirical Bayes normalization approach and clarifying assumptions of existing scaling methods.

The paper evaluates 16S rRNA marker gene analysis pipelines using a mixture dataset of human stool samples, finding DADA2 better captures specific features than Mothur and QIIME.