Unsupervised Two-Way Clustering of Metagenomic Sequences

Table 2

Performance of Poisson mixture model on datasets for different values of 𝐿 and word length of 5. Here, N.W.G stands for no word grouping. The maximum accuracy achieved is in bold. Each dataset contains 50,000 reads of length 500 bp.

Species	𝐿=5	𝐿=10	𝐿=30	𝐿=50	N.W.G
B. anthracis CI chromosome, B. halodurans C-125	90.61	91.53	50.31	91.2	50.32
H. pylori 26695, S. pneumoniae 70585	98.6	98.79	98. 73	98.71	98.76
B. subtilis subsp. spizizenii str., L. lactis subsp.	89.96	90.34	90.62	90.53	50.47