Erin M Hanna1, Timothy J Hamp2, Iain H McKillop3, Farah Bahrani-Mougeot4, John B Martinie1, James M Horton5, David Sindram1, Raad Z Gharaibeh6, Anthony A Fodor2, David A Iannitti1. 1. Department of Surgery, Carolinas Medical Center, Charlotte, North Carolina. 2. Department of Bioinformatics and Genomics, UNC Charlotte, Charlotte, North Carolina. 3. Department of Surgery, Carolinas Medical Center, Charlotte, North Carolina. Electronic address: iain.mckillop@carolinashealthcare.org. 4. Department of Oral Medicine, Carolinas Medical Center, Charlotte, North Carolina. 5. Department of Internal Medicine, Carolinas Medical Center, Charlotte, North Carolina. 6. Department of Bioinformatics and Genomics, UNC Charlotte, Charlotte, North Carolina; Department of Bioinformatics and Genomics, Bioinformatics Services Division, UNC Charlotte, Kannapolis, North Carolina.
Abstract
BACKGROUND: Infected necrotizing pancreatitis is associated with significant morbidity and mortality. Peripancreatic fluid cultures may fail to identify all the infecting organisms. The aim of this study was to compare the bacterial biome of peripancreatic fluid from infected necrotizing pancreatitis patients using 16S ribosomal RNA (rRNA) DNA deep sequencing and quantitative polymerase chain reaction (qPCR) targeting the 16S rRNA gene versus standard laboratory culture. MATERIALS AND METHODS: Peripancreatic fluid was collected during operative or radiologic intervention and samples sent for culture. In parallel, microbial DNA was extracted, qPCR targeting the 16S rRNA gene and 16S rRNA PCR amplification followed by Illumina deep sequencing were performed. RESULTS: Using culture techniques, the bacterial strains most frequently identified were gram-negative rods (Escherichia coli, Klebsiella pneumoniae) and Enterococcus. Samples in which culture results were negative had copy numbers of the 16S rRNA gene close to background in qPCR analysis. For samples with high bacterial load, sequencing results were in some cases in good agreement with culture data, whereas in others there were disagreements, likely due to differences in taxonomic classification, cultivability, and differing susceptibility to background contamination. Sequencing results appeared generally unreliable in cases of negative culture where little microbial DNA was input into qPCR sequencing reactions. CONCLUSIONS: Both sequencing and culture data display their own sources of bias and potential error. Consideration of data from multiple techniques will yield a more accurate view of bacterial infections than can be achieved by any single technique.
BACKGROUND:Infected necrotizing pancreatitis is associated with significant morbidity and mortality. Peripancreatic fluid cultures may fail to identify all the infecting organisms. The aim of this study was to compare the bacterial biome of peripancreatic fluid from infected necrotizing pancreatitispatients using 16S ribosomal RNA (rRNA) DNA deep sequencing and quantitative polymerase chain reaction (qPCR) targeting the 16S rRNA gene versus standard laboratory culture. MATERIALS AND METHODS: Peripancreatic fluid was collected during operative or radiologic intervention and samples sent for culture. In parallel, microbial DNA was extracted, qPCR targeting the 16S rRNA gene and 16S rRNA PCR amplification followed by Illumina deep sequencing were performed. RESULTS: Using culture techniques, the bacterial strains most frequently identified were gram-negative rods (Escherichia coli, Klebsiella pneumoniae) and Enterococcus. Samples in which culture results were negative had copy numbers of the 16S rRNA gene close to background in qPCR analysis. For samples with high bacterial load, sequencing results were in some cases in good agreement with culture data, whereas in others there were disagreements, likely due to differences in taxonomic classification, cultivability, and differing susceptibility to background contamination. Sequencing results appeared generally unreliable in cases of negative culture where little microbial DNA was input into qPCR sequencing reactions. CONCLUSIONS: Both sequencing and culture data display their own sources of bias and potential error. Consideration of data from multiple techniques will yield a more accurate view of bacterial infections than can be achieved by any single technique.