Fig. 2

The processing pipeline for classifying metagenomic reads to one of the constructed classes. D_H—k-mer database for the human reference sequence; \(\left \{D^{0}_{1}, D^{0}_{2}, \dots, D^{0}_{N}\right \}\)—k-mer databases from the original datasets for each of N classes; {D₁,D₂,…,D_N}—k-mer databases after subtracted D_H for each of N classes; R_i—an ith read from a query sample which includes \(\mathfrak {q}\) reads; Ξ_RiDj—a result of matching a jth read to the ith class (match rate score); x—one of the constructed classes; each blue block indicates data stored in a separate file