ldiao/Gimp: Good-Turing and empirical Bayes pseudocounts for sparse count matrices

16S rRNA sequence data is oftentimes extremely sparse, making it difficult to normalize properly for differential abundance analyses with current RNA-Seq methods, such as DESeq, edgeR, and metagenomeSeq. The count adjustment methods included here can be used to ameliorate these effects. The Good-Turing adjustment is derived from from Good-Turing frequency estimation methods. Code to generate simulated test data according to the Dirichlet-multinomial model from two different sets of parameters is also included. Further information on the methods and data can be found *IN THIS PAPER*.

install.packages("remotes")
remotes::install_github("ldiao/Gimp")