“Compositional uncertainty should not be ignored in high-throughput sequencing data analysis” (2016) Austrian Journal of Statistics, 45(4), pp. 73–87. doi:10.17713/ajs.v45i4.122.