The Inter-Laboratory Comparisons Of Metagenomic Next-Generation Sequencing On Multiple Samples

Infectious diseases are still a global public health problem that threatens human health.Culture-based detection of pathogens is the gold standard for infectious disease diagnostics but has the limitation of low sensitivity.In recent years,rapid pathogen diagnosis based on second-generation sequencing is widely applied in clinical settings because of its high throughput and sensitivity.Metagenomic next-generation sequencing(mNGS)technology refers to the unbiased detection of multiple pathogenic microorganisms by high-throughput birdshot sequencing of the entire biological genome in clinical samples.However,differences in mNGS library construction and analysis processes between laboratories interfere with the consistency of reporting and diagnosis of disease.This study aimed to assess the inter-laboratory variation in mNGS testing using real clinical samples,but also to analyze the key factors affecting mNGS results,and to provide a reference for the selection of appropriate testing conditions and report interpretation.1 Objectives1.1 To explore the differences in mNGS assays between laboratories and compare the differences between mNGS and traditional methods;1.2 To explore the specific differences in sequencing results between laboratories and what are the specific factors that contribute to the differences;1.3 To provide guidance on the clinical application of mNGS.2 MethodsIn this study,100 patients with suspected upper respiratory tract infection and 24 patients with suspected bloodstream infection were included,and their broncho alveolar lavage fluid(BALF)and blood samples were collected.All samples were tested for infectious agents by traditional methods and distributed simultaneously to three laboratories for high-throughput macrogenic and macrotranscriptomic testing.Blood samples were tested for macrogenome only.This experiment was primarily based on descriptive analysis.Firstly,we evaluated the specificity and sensitivity of common pathogens mNGS in each laboratory based on the traditional test results.For microorganisms outside the range of traditional detection,we analyzed the consistency of microbial species detected by mNGS between laboratories.In addition,we extracted the original sequencing data from half of the samples and compared the influence of different laboratory database construction methods,sequencing read length,and analysis process differences on the consistency of the results of mNGS results.At the same time,we analyzed macro transcriptome sequencing reports of samples to evaluate the contribution of RNA-based sequencing approaches to rapid pathogen diagnosis.Finally,we used patients’ inflammation-related clinical indicators combined with the number of pathogen sequences to evaluate the influence of patient status and infection degree on the detection of mNGS pathogens.3 ResultsCompared with the results of traditional examination,the total mean sensitivity of detection of common pathogens in the alveolar lavage fluid samples was 64.6%.The total specificity was 96.6%.While in blood samples,the total mean value of sensitivity for detection of common pathogens was 29.1%.The specificity was 100%.Negative concordance rates among the three laboratories were located between 6.8%and 60%.A total of 125 pathogens were reported at the genus level in the mNGS assays ofthe three laboratories,of which only 4 microorganisms could achieve perfect concordance in the results of all samples.Nine microorganisms were detected with a concordance rate higher than 50%,and as many as 90 microorganisms had a concordance rate of 0.According to the sequence analysis,all three labs had a high proportion of Q20 and Q30 reads;human-derived sequences accounted for a higher proportion,and the effect of de-hosting nucleic acid treatment could be improved;different comparison methods would affect species detection and identification.Meanwhile,we found that data with reads longer than 70bp were better classified than those below 50bp.Macro-transcriptional sequencing results from various laboratories showed that only 9%,18%and 6%of samples could increase the positive rate of detection of potential pathogens by this sequencing method..The degree of infection had a positive convergent effect on the mNGS testing,but there were no significant inter-laboratory differences.4 ConclusionAll laboratories have high specificity and sensitivity in the detection of common pathogens.The detection concordance rate was significantly lower in specimens that were negative for conventional testing and correlated with specimen type.There is significant variation in the detection and judgment of pathogens among laboratories,and differences may arise from different library construction and comparison methods,which reduce the confidence of individual laboratory results.mNGS assays play a complementary role to traditional tests,especially in the detection of pathogens in mild infections,but still cannot completely replace them.Macro-transcriptomic assays have greater efficacy in detecting RNA pathogens only.The degree of infection had a positive effect on mNGS pathogen detection,with little difference between laboratories.