A systematic sequencing-based approach for microbial contaminant detection and functional inference

BMC Biology

  ^{1    2    3    3    4    5} 

[1] 0000 0001 0720 6587, grid.410818.4, Institute of Advanced Biomedical Engineering and Science, Tokyo Women’s Medical University, 162-8666, Tokyo, Japan;0000 0001 1014 9130, grid.265073.5, Department of Periodontology, Graduate School of Medical and Dental Sciences, Tokyo Medical and Dental University, 113-8549, Tokyo, Japan;0000 0001 0720 6587, grid.410818.4, Institute of Advanced Biomedical Engineering and Science, Tokyo Women’s Medical University, 162-8666, Tokyo, Japan;0000 0004 0372 2359, grid.411238.d, Division of Periodontology, Department of Oral Function, Kyushu Dental University, 803-8580, Fukuoka, Japan;0000 0001 2151 536X, grid.26999.3d, Department of Computational Biology and Medical Sciences, Graduate School of Frontier Sciences, The University of Tokyo, 277-8568, Kashiwa, Japan;0000 0001 2151 536X, grid.26999.3d, Human Genome Center, The Institute of Medical Science, The University of Tokyo, 108-8693, Tokyo, Japan;0000 0001 2151 536X, grid.26999.3d, Human Genome Center, The Institute of Medical Science, The University of Tokyo, 108-8693, Tokyo, Japan;0000 0001 2151 536X, grid.26999.3d, Department of Computational Biology and Medical Sciences, Graduate School of Frontier Sciences, The University of Tokyo, 277-8568, Kashiwa, Japan;

关键词: Contamination;    Mycoplasma;    Host-microbe interaction;    Next-generation sequencing;    Non-negative matrix factorization;   

DOI  :  10.1186/s12915-019-0690-0

来源: publisher

【 摘 要 】

BackgroundMicrobial contamination poses a major difficulty for successful data analysis in biological and biomedical research. Computational approaches utilizing next-generation sequencing (NGS) data offer promising diagnostics to assess the presence of contaminants. However, as host cells are often contaminated by multiple microorganisms, these approaches require careful attention to intra- and interspecies sequence similarities, which have not yet been fully addressed.ResultsWe present a computational approach that rigorously investigates the genomic origins of sequenced reads, including those mapped to multiple species that have been discarded in previous studies. Through the analysis of large-scale synthetic and public NGS samples, we estimate that 1000–100,000 contaminating microbial reads are detected per million host reads sequenced by RNA-seq. The microbe catalog we established included Cutibacterium as a prevalent contaminant, suggesting that contamination mostly originates from the laboratory environment. Importantly, by applying a systematic method to infer the functional impact of contamination, we revealed that host-contaminant interactions cause profound changes in the host molecular landscapes, as exemplified by changes in inflammatory and apoptotic pathways during Mycoplasma infection of lymphoma cells.ConclusionsWe provide a computational method for profiling microbial contamination on NGS data and suggest that sources of contamination in laboratory reagents and the experimental environment alter the molecular landscape of host cells leading to phenotypic changes. These findings reinforce the concept that precise determination of the origins and functional impacts of contamination is imperative for quality research and illustrate the usefulness of the proposed approach to comprehensively characterize contamination landscapes.

【 授权许可】

CC BY