Depletion of Human DNA in Spiked Clinical Specimens for Improvement of Sensitivity of Pathogen Detection by Next-Generation Sequencing

Abstract

Next-generation sequencing (NGS) technology has shown promise for the detection of human pathogens from clinical samples. However, one of the major obstacles to the use of NGS in diagnostic microbiology is the low ratio of pathogen DNA to human DNA in most clinical specimens. In this study, we aimed to develop a specimen-processing protocol to remove human DNA and enrich specimens for bacterial and viral DNA for shotgun metagenomic sequencing. Cerebrospinal fluid (CSF) and nasopharyngeal aspirate (NPA) specimens, spiked with control bacterial and viral pathogens, were processed using either a commercially available kit (MolYsis) or various detergents followed by DNase prior to the extraction of DNA. Relative quantities of human DNA and pathogen DNA were determined by real-time PCR. The MolYsis kit did not improve the pathogen-to-human DNA ratio, but significant reductions (>95%;P< 0.001) in human DNA with minimal effect on pathogen DNA were achieved in samples that were treated with 0.025% saponin, a nonionic surfactant. Specimen preprocessing significantly decreased NGS reads mapped to the human genome (P< 0.05) and improved the sensitivity of pathogen detection (P< 0.01), with a 20- to 650-fold increase in the ratio of microbial reads to human reads. Preprocessing also permitted the detection of pathogens that were undetectable in the unprocessed samples. Our results demonstrate a simple method for the reduction of background human DNA for metagenomic detection for a broad range of pathogens in clinical samples.

Copyright © 2016, American Society for Microbiology. All Rights Reserved.

Figures

FIG 1

Processing of clinical specimens using MolYsis reagents. CSF (A) or NPA (B) specimens were spiked with S. pneumoniae and influenza A virus and processed by MolYsis method I and MolYsis method II, as described in the Materials and Methods. Nucleic acids extracted from processed or unprocessed samples were analyzed by real-time PCR assays for β-actin DNA and β-2-microglobulin RNA for human DNA and RNA, respectively, and pathogen-specific targets, as described in Table 1. The percentage of DNA/RNA was obtained from fold changes calculated from CT values, with respect to CT values for unprocessed samples. The data are the average of data obtained from 3 independent experiments using a total of 3 CSF specimens and 3 NPA specimens. Error bars are the standard error of the mean; statistical significance was calculated by the pairwise Student's t test (two-tailed).

FIG 2

Processing of clinical specimens using various detergents for selective lysis of human cells. NPA specimens were spiked with S. pneumoniae, HSV2, and adenovirus and treated with saponin, Triton X-100, Tween 20, CHAPS buffer, or MolZym CM buffer to final concentrations indicated in the figure, followed by DNase treatment (see Materials and Methods). Nucleic acids extracted from processed or unprocessed samples were analyzed by real-time PCR assays for β-actin DNA for human DNA and pathogen-specific targets, as described in Table 1. The percentage of DNA was obtained from fold changes calculated from CT values, with respect to CT values for unprocessed samples. The data are the average of data obtained from 3 independent experiments using a total of 4 NPA specimens. Error bars are the standard error of the mean; statistical significance was calculated by the pairwise Student's t test (two-tailed). *, P < 0.05; **, P < 0.01; ***, P < 0.001.

FIG 3

Processing of clinical specimens using saponin at different concentrations for selective lysis of human cells. CSF specimens were spiked with S. pneumoniae, H. influenzae, E. coli, HSV2, and adenovirus (A) and NPA specimens spiked with S. pneumoniae, H. influenzae, B. pertussis, and adenovirus (B) were treated with saponin and Triton X-100 to final concentrations indicated in the figure, followed by DNase treatment (see Materials and Methods). Nucleic acids extracted from processed or unprocessed samples were analyzed by real-time PCR assays for β-actin DNA for human DNA and pathogen-specific targets, as described in Table 1. The percentage of DNA was obtained from fold changes calculated from CT values, with respect to CT values for unprocessed samples. The data are the average of data obtained from 3 independent experiments using a total of 4 CSF specimens and 4 NPA specimens. Error bars are the standard error of the mean; statistical significance was calculated by the pairwise Student's t test (two-tailed); *, P < 0.05; **, P < 0.01; ***, P < 0.001.

FIG 4

Analysis of clinical specimens processed with 0.025% saponin and DNase by PCR and NGS. CSF specimens (n = 3) spiked with S. pneumoniae, N. meningitidis, H. influenzae, E. coli, S. agalactiae, HSV2, and adenovirus and NPA specimens (n = 3) spiked with S. pneumoniae, H. influenzae, B. pertussis, HSV2, and adenovirus were treated with saponin to a final concentration of 0.025%, followed by DNase treatment (see Materials and Methods). (A) Nucleic acids extracted from processed or unprocessed samples were analyzed by real-time PCR assays for β-actin DNA for human DNA, and pathogen-specific targets, as described in Table 1. The percentage of DNA was obtained from fold changes calculated from CT values, with respect to CT values for unprocessed samples. Data for all spiked pathogens were averaged; error bars are the standard error of the mean. (B) Nucleic acids extracted from processed or unprocessed samples were analyzed by NGS on Illumina MiSeq. Sequencing libraries were constructed from 1 ng of DNA from specimens using Nextera XT. Raw data were filtered for quality and mapped to human genome and pathogen genomes, as described in the Materials and Methods. Specimens were processed and analyzed independently, as they were available.