HMP-package: Hypothesis Testing and Power Calculations for Comparing...

Description Details Author(s) References


This package provides tools for Generating data matrices following Multinomial and Dirichlet-Multinomial distributions, Computing the following test-statistics and their corresponding p-values, and Computing the power and size of the tests described above using Monte-Carlo simulations.


\textbf{Hypothesis Test} \textbf{Test Statistics Function} \textbf{Power Calculation Function}
2+ Sample Means w/ Reference Vector Xmc.sevsample MC.Xmc.statistics
1 Sample Mean w/ Reference Vector Xsc.onesample MC.Xsc.statistics
2+ Sample Means w/o Reference Vector Xmcupo.sevsample MC.Xmcupo.statistics
2+ Sample Overdispersions Xoc.sevsample MC.Xoc.statistics
2+ Sample DM-Distribution Xdc.sevsample MC.Xdc.statistics
Multinomial vs DM C.alpha.multinomial MC.ZT.statistics

In addition to hypothesis testing and power calculations you can:

  1. Perform basic data management to exclude samples with fewer than pre-specified number of reads, collapse rare taxa and order the taxa by frequency. This is useful to exclude failed samples (i.e. samples with very few reads) - Data.filter

  2. Plot your data -

  3. Generate random sample of Dirichlet Multinomial data with pre-specified parameters - Dirichlet.multinomial

Note: Thought the description of the functions refer its application to ranked abundance distributions (RAD) data, every function is also applicable to model species abundance data. See references for a discussion and application to both type of ecological data.


Patricio S. La Rosa, Elena Deych, Berkley Shands, Sharina Carter, Dake Yang, William D. Shannon


La Rosa PS, Brooks JP, Deych E, Boone EL, Edwards DJ, et al. (2012) Hypothesis Testing and Power Calculations for Taxonomic-Based Human Microbiome Data. PLoS ONE 7(12): e52078. doi:10.1371/journal.pone.0052078

Yang D, Johnson J, Zhou X, Deych E, et al. (2019) Microbiome Recursive Partitioning. Currently Under Review.

HMP documentation built on Aug. 31, 2019, 5:05 p.m.