Metagenomic Sequencing Detects Respiratory Pathogens in Hematopoietic Cellular Transplant Patients

C Langelier; MS Zinter; K Kalantar; GA Yanik; S Christenson; B Odonovan; C White; M Wilson; A Sapru; CC Dvorak; S Miller; CY Chiu; JL DeRisi

The respiratory tract is not a sterile environment, and as such, we devised a ranking score to aid in the stratification of the inherently complex alignment data. This metric involved multiplying the number of reads aligned per million reads sequenced by the sum of the nt and nr Z-scores for each microbial genus identified in each patient: score = rpM nt x (Z nt + Z nr ). To further improve the specificity of microbial assignments, we required 1) microbes to be detected in both nucleotide (nt) and protein (nr) database alignments; 2) microbes to have a Z nt and Z nr > zero and 3) bacteria and fungi detected by RNAseq to also have detectable genomic sequence by DNAseq. In addition, we permitted DNA viruses to be detected on only DNAseq and RNA viruses to be detected on only RNAseq. For each patient, any microbes that passed these requirements were then processed using the following decision tree, summarized in Figure 1 .