Analytical Validation of a Highly Accurate and Reliable Next-Generation Sequencing-Based Urine Assay

Mara Couto-Rodriguez, David C Danko, Heather L Wells, Sol Rey, Xavier Jirau Serrano, John Papciak, P Ford Combs, Gabor Fidler, Christopher E. Mason, Caitlin Otto, Niamh B. O’Hara, Dorottya Nagy-Szakal
doi: https://doi.org/10.1101/2024.04.05.24305286

ABSTRACT

Culture is currently the gold standard for diagnosis of urinary tract infections (UTIs); however, it has poor sensitivity detecting urogenital pathogens, especially if patients have already initiated antimicrobial therapy, or have an infection from an organism that is not commonly cultured. False negative urine culture results can lead to the inappropriate use of antimicrobial therapies or to the progression to urosepsis in high-risk patients. Though not commonly applied to urine in a clinical setting, Next-generation sequencing (NGS)-based metagenomics offer a solution as a precision diagnostic. We developed and validated BIOTIA-ID, a clinical-grade NGS-based diagnostic pipeline for the detection and identification of pathogens in urine specimens. Remnant clinical urine specimens, and contrived sterile urine spiked with common UTI pathogens, were processed with our end-to-end assay including extraction, metagenomic library preparation and Illumina NextSeq 550 sequencing. We trained and applied a bioinformatic pipeline that uses machine learning (ML) to identify pathogens. Internal controls and other quality control measures were incorporated into the process to provide rigorous and standardized results. The assay was tested on 1,470 urine specimens and achieved 99.92% sensitivity, 99.95% specificity and a limit of detection (LoD) of <25,000 CFU/mL and <5,000 CFU/mL in bacteria and fungi, respectively. Discordant results were reconciled with additional testing by target-specific qPCR or 16S Sanger sequencing; 87% of the NGS results were ultimately determined to be the correct result. Overall, these data demonstrate that BIOTIA-ID is a highly accurate clinical-grade diagnostic tool with notable advantages over current culture- based diagnostics.

Conflict of Interest Statement MCR, DCD, HLW, SR, XJS, JP, PFC, GF, CEM, CO, NBO and DNS are employees at Biotia, Inc.

INTRODUCTION

Annually, 11 million people in the United States and 404 million people worldwide are diagnosed with a urinary tract infection (UTI) (1, 2). Immunocompromised patients and patients with urological conditions are at higher risk to develop complicated and/or recurrent UTI (cUTI/rUTI) as well as to progress to urosepsis, resulting in higher morbidity and mortality and increased cost of care(3–8). In hospitalized patients, UTIs are associated with 2.3% of the mortality rate and an estimated annual cost of $340 to $450 million in the United States alone(1). Approximately 30% of septic patients have an infection originating from the urogenital tract, with a multinational study revealing that 12% of nosocomial UTI patients progressed to urosepsis (7, 9). Rapidly identifying UTI pathogens in high-risk patients is crucial to administering an appropriate treatment and minimizing unnecessary broad-spectrum antibiotic use.

Urine culture is the standard of care (SOC) for the detection and identification of pathogens causing UTIs. However, culture-dependent approaches are focused on an Enterobacterales-centric paradigm which is based on assumptions that urine is sterile (10–14), disregards mixed infections as contamination (15–17), and is based on clinical studies employing heterogeneous definitions and thresholds for a UTI (18, 19). Culture has several limitations, including: (1) an inability to identify hard-to-grow microbial organisms, such as anaerobes and fungi; (2) challenges in identifying rare pathogens or organisms not routinely cultured; and (3) risk of false negative results for patients being treated with antibiotics (20–23). Long diagnostic turnaround time (TAT) or inconclusive results could lead to the empiric treatment of suspected infections with broad-spectrum antibiotics that are often inappropriate and may contribute to increased rates of drug resistance (24, 25).

The emergence of new molecular and culture-based techniques has underscored the drawbacks of traditional culture and challenged our understanding of UTIs, raising questions around what is considered a urogenital pathogen and the role of mixed bacterial communities in UTIs(12, 20, 26–32). Genomics-based assays enable direct specimen analysis without the need for culturing by sequencing genetic material, which is analyzed for the presence of pathogenic organisms through detection of their DNA. By comparing detected genetic sequences to a comprehensive microbial genome database, genomics-based tests can accurately and rapidly identify pathogens in cUTI samples, which may have atypical pathogens, multiple co-infecting organisms, and complex drug resistance profiles. Concerns have been raised about the sensitivity of molecular approaches to differentiate between opportunistic and commensal pathogen detection and differentiation of colonization versus infection (30, 33–36). However, careful development of genomics-based assays that center on extensive validation leveraging culture, molecular and bioinformatics techniques can increase the utility of this methodology in routine clinical practice, thus offering promising supplemental testing method to the current gold standard UTI testing.

Here, we developed, optimized, and clinically validated a (NGS)-based urine assay, BIOTIA-ID, that has high diagnostic sensitivity and specificity and provides comprehensive detection of urogenital pathogens within a single test.

MATERIALS AND METHODS

Clinical Validation Strategy

All specimens were processed with our version-controlled BIOTIA-ID Urine NGS Assay laboratory protocol, analytical pipeline, and reference microbial database in Biotia’s CLIA- certified laboratory. The assay and clinical validation were designed based on New York State Department of Health validation guidelines for submission of a Laboratory Developed Test for bacteriology and mycology nucleic acid amplification assays. An overview of the assay workflow and clinical validation design is illustrated in Figure 1.

Figure 1.
Figure 1.

(A) An overview of the BIOTIA-ID Urine NGS Assay workflow, consisting of sample collection (urine specimens in standard UTT), nucleic acid extraction, library preparation and sequencing using the Illumina platform, BIOTIA-DX metagenomic analysis, multiple decision tree-based machine learning modeling and generation of a clinical report with urogenital pathogens detected. (B) Analytical validation including specificity, sensitivity and in silico studies evaluated the clinical performance characteristics of the assay. The clinical validation established the accuracy of urogenital pathogen detection in clinical specimens with additional comparator studies using qPCR and Sanger sequencing in cases where culture and NGS results disagreed. Furthermore, we completed an inter/intra assay reproducibility study to show the validity and consistency of our results.

Clinical Specimens and Reference Materials

Deidentified urine specimens in urine transport tubes (UTT) were collected as residual samples after routine clinical testing from different microbiology reference laboratories and were provided with the culture results for each specimen following their standard operating procedures to report pathogens. Samples were collected and processed under an institutional review board (Advarra Pro00038083) without demographic information. A collection of clinical and reference isolates (ATCC, Zeptometrix) was also used for the clinical validation including bacteria, fungi, viruses, and parasites (listed in Table 1). Reference and clinical microbial isolates used in the contrived specimens were inoculated and grown in Tryptic Soy Agar with 5% Sheep Blood at 37°C for 24-48 hours.

View this table:
Table 1.

Reference organisms and clinical isolates tested and validated on the specificity study. Whole organisms of urogenital pathogens detected by the assay, related organisms, and other organisms that can be present in urine specimens were tested by spiking microbial cells into a negative urine matrix (50,000 CFU/mL for bacteria and 25,000 CFU/mL for fungi). A total of 69 microbial species (bacteria, fungi, viruses and parasites) and 154 strains were evaluated. (*) Indicates organisms tested with BIOTIA-ID assay but not reported as a key urogenital pathogen.

Clinical-Grade Metagenomic Sequencing for Infectious Disease Diagnostics

BIOTIA-ID Urine NGS Assay. Genomic DNA (gDNA) extraction from 2 mL of urine was performed with QIACube-MDx and yields were quantified with the Qubit Flex. Normalized gDNA was spiked with 5% internal positive control (IPC) and processed with the Illumina DNA prep library preparation kit. Libraries were quality checked for size and concentration with Tapestation 4200 and Qubit Flex, respectively. Libraries were pooled in equimolar concentrations and sequenced on an Illumina NextSeq 550 platform.

Assay Controls. Each NGS run includes three external controls and one internal control, which are processed with each batch of samples. The external controls consist of a no template control (NTC, nuclease free water), positive control (PC, Zymobiomics) and negative extraction control (NEC, negative urine matrix). The IPC is a proprietary gDNA that is spiked into each sample at 5%.

BIOTIA-DX Software. After sequencing, the FastQ files were first subjected to a filtering step which removed low-quality reads and sequencing adapters. Detection of the internal positive control was performed for each sample as a confirmation of overall sequencing quality. Human sequence data was then removed from downstream analysis. The remaining reads were first compared against a large database of microbial genomes in a coarse classification step using a commonly used metagenomics tool, kraken2 (37). We use a custom kraken2 microbial database within BIOTIA-DX that includes curated, high-quality genetic sequences from bacterial, fungal, viral, and non-fungal eukaryotic parasitic organisms. In total, more than 7,000 organisms are represented in the database, with diverse representation of microbial species and strains.

Despite any degree of database curation, metagenomics tools often result in spurious, false positive hits, which has historically made it difficult to use NGS in a clinical diagnostic setting(38–41). To distinguish between clinically relevant infections, off-target spurious hits by the metagenomics software, or low levels of cross-contamination, we designed a classifier using ML that estimates the probability that a microbial organism is present in a sample. Targets exceeding predefined kraken2 thresholds in the coarse classification step were tested using this ML classifier in a fine classification step. Finally, organisms which exceeded a pre-defined probability threshold from the classifier were called as present in the sample.

Pre-Validation (BIOTIA-DX Training Set)

To ensure clinical-grade diagnostic accuracy of these detection thresholds, we trained our ML classifier to recognize clinical-level infections using a training set of clinical samples with known and orthogonally validated infectious organisms. We sequenced and tested a pre-validation sample set of 114 clinical urine samples with known organisms identified by culture, and in cases where our algorithm results differed from culture results, we used PCR testing as a comparator assay. Results from PCR tests were then used to update the list of known organisms in the pre-validation set and re-train the ML classifier with the updated results. The classifier uses a variety of bioinformatic features chosen specifically to distinguish true infections from false positives. These features were calculated for each organism being tested and included statistics such as the depth and evenness of genome coverage, read identity to reference genomes, and the quality of assembled contigs.

Analytical Validation

Specificity. Reference organisms and clinical isolates were tested and validated to establish the specificity of the assay. Whole organisms of urogenital pathogens detected by the assay, genetically related organisms, and other organisms that can be present in urine specimens were tested by spiking microbial cells into negative urine matrix. A total of 69 microbial species (bacteria, fungi, viruses, and parasites) and 154 strains were evaluated (Table 1). Microbial isolates were spiked into a negative urine matrix at concentrations of 25,000-50,000 CFU/mL and processed with the assay.

Analytical Sensitivity. The analytical sensitivity was assessed in 16 of the most prevalent urogenital pathogens, by spiking reference whole organisms into negative urine matrix (Bacteria: Acinetobacter baumanii, Bacteroides fragilis, Escherichia coli, Enterococcus faecalis, Gardnerella vaginalis, Klebsiella pneumoniae, Prevotella spp., Proteus mirabilis, Pseudomonas aeruginosa and Staphylococcus aureus; Fungi: Candida albicans, Candida auris, Candida glabrata, Candida krusei, Candida parapsilosis, Candida tropicalis). Serial dilutions of each pathogen were spiked in triplicate followed by gDNA extraction. The limit of detection was defined when the target organism is detected in ≥95% of the samples.

In silico Analytical Specificity. To assess the analytical specificity of our pipeline, we used a total of 12,264 RefSeq genomes from 5,847 unique bacteria, eukaryotic organisms, and viruses, to simulate sequencing files with 100,000 reads each. A random mutation rate of 0.5% was applied during simulation to introduce a reasonable degree of variability. Simulated read files were then run through BIOTIA-DX in an identical manner to sequenced laboratory samples and tested for key organisms to assess the potential for organisms outside our laboratory sample set to generate false positive results for key organisms.

Inter/Intra-Reproducibility. For assessment of the reproducibility in bacterial detection, five different pools containing one clinical strain each of E. coli, E. faecalis, K. pneumoniae, P. mirabilis and S. aureus were tested. Each pool had different strains for each species (altogether 25 different strains) and were spiked into a negative urine matrix at a concentration of 25,000 CFU/mL per analyte. Each pool was then tested daily for three consecutive days in three technical replicates for a total of 15 replicates per pool), 45 replicates per analyte and 6 negative samples. A smaller follow up study sought to evaluate reproducibility of bacterial detection at different concentrations 100,000 CFU/mL (high), 25,000 and 20,000 CFU/mL (medium) and 15,000 CFU/mL (low). Two bacterial pools per concentration were tested in triplicate across three different days for a total of 9 replicates per concentration per analyte. A similar experimental design was performed to assess fungal reproducibility of ten fungal analytes. Five different pools each containing one different strain of C. albicans, C. auris, C. glabrata, C. krusei, C. parapsilosis, and C. tropicalis and a sixth pool with one strain of C. dubliniensis, C. guilliermondii, C. kefyr and C. lusitaniae were tested in triplicate across three different days. A total of 15 replicates per pool, 45 replicates per analyte and 27 negative samples were evaluated.

Clinical validation

Accuracy. We tested a combination of urine clinical specimens and contrived samples (whole organism microbial reference strains and clinical isolates spiked into negative clinical matrix) to assess accuracy. Deidentified, remnant clinical urine samples were obtained from a reference laboratory and stored at -80°C until processing with the assay. Urine clinical samples were collected based on the pathogens diagnosed by culture. At least 30 specimens each with the most common urogenital pathogens (E. coli, E. faecalis, K. pneumoniae, P. mirabilis and S. aureus) were collected. When fewer than 30 samples for a given pathogen were available, contrived specimens were used to bring the total number of samples for that target to 30 (a total of 335 samples; 167 clinical specimens and 168 contrived). Contrived samples were used for S. aureus (n=30), P. mirabilis (n=9), K. pneumoniae (n=8), C. albicans (n=33), C. auris (n=30), C. glabrata (n=30), C. krusei (n=33), C. parapsilosis (n=34), C. tropicalis (n=30) with a spike in concentration close to the LoD (∼2.5 - 5× LoD). In addition, 35 specimens negative by culture were tested.

Comparator Testing. gDNA from urine specimens was used to perform confirmatory testing of clinical samples yielding results in which NGS and culture findings did not agree. Discrepant results were adjudicated either with qPCR or Sanger Sequencing. Taxa- or group-specific molecular assays were used to amplify analytes of interest (primers and references are listed in Supplemental Table 1). PCR amplicons were sent for Sanger sequencing and consensus sequences were used to determine percent identity of the organism detected in the clinical specimen. A total of 8 different target-specific qPCR assays and 8 Sanger sequencing assays were used to test 123 discrepant taxa and 19 concordant taxa.

Code Availability

The BIOTIA-DX software used is described in the Methods section. The software leverages the following openly available tools: fastp v0.23.3, bowtie2 v2.5.1, minimap2 v2.22, SNAP aligner 2.0.3, samtools v1.6, and kraken2 v2.1.3. The ML classifier is a proprietary portion of the code(37, 42–46).

Data Availability

The data supporting the study findings are available from the corresponding author on request. Sequencing data that support the finding of this study (with human reads removed) have been deposited in GeoSeeq, a publicly available metagenomic platform, and available upon request.

RESULTS

We developed an end-to-end urine clinical diagnostic metagenomics assay to detect urogenital pathogens and validated its performance on 1,470 samples sequenced in 65 sequencing runs (Figure 1). Clinical and analytical validation experiments were conducted with clinical and contrived specimens to assess specificity, sensitivity, BIOTIA-DX in silico performance, reproducibility, and accuracy (Figure 1B).

Microbial read counts, IPC detection, and external positive (PC) and negative controls (NEC and NTC) sequencing performance was assessed. Validation samples averaged 7.9M raw reads and 4.7M microbial reads after human removal (Figure 2A). The PC generated similar read depths as testing samples, while the NEC and NTC had reads < 750K. The IPC was detected within expected range (1-5% relative abundance), except in the negative controls (75% relative abundance) (Figure 2B).

Figure 2.
Figure 2.

The assay performance of specimens processed in the analytical (specificity, sensitivity, and reproducibility) and clinical validation (accuracy) studies was defined by evaluating the internal positive control (IPC) and the number of microbial reads. (A) The IPC was detected in all specimens and positive controls (PC) at an expected range (1-5% abundance). The IPC reads represented the majority of the reads detected in negative extraction controls (NEC) and negative template controls (NTC). (B) Violin plots depicting the number of microbial reads obtained in each study and controls. As expected, the clinical and contrived specimens processed through the studies and the PCs generated similar microbial read depth (average 7.9M of microbial reads), while the NEC and NTC had microbial reads below 750k. Violin plots show IPC abundance and microbial read distribution for all samples and controls processed in each study: Specificity (light blue, n=160); Sensitivity (navy blue, n=709); Reproducibility (light pink, n= 267); Accuracy (gray, n= 325), PC (dark pink, n=65), NEC (light gray, n=130), NTC (cerulean blue, n=130)

Key performance characteristics of BIOTIA-ID. SPECIFICITY

The similarity among different microbial genomes, the diversity of clinical isolates from reference genomes and contaminating DNA fragments impacts the analytical sensitivity of a metagenomic test. Urogenital pathogens detected by the assay and genetically related organisms were processed with BIOTIA-ID (Table 1), yielding a sensitivity and specificity of 100% and 99.96% for bacterial targets and 100% sensitivity and specificity for fungal targets. During assay development, several taxa not extensively represented in the clinical training set generated false positive and false negative results leading us to process 64 additional species in 80 contrived urine samples to feed more training data to the model. This approach coupled with hyperparameter tuning of the model ensued the increased sensitivity and specificity reported above.

SENSITIVITY

The analytical sensitivity studies established the lowest concentration of pathogen reliably detected by our assay. Ten bacterial and six fungal species were tested as 6 replicates per concentration, per analyte evaluated (Figure 3A). A total of 709 contrived specimens were processed resulting in an overall LoD of <25,000 CFU/mL for bacteria and <5,000 CFU/mL for fungi. Specifically, the LoDs established were 1,000 CFU/mL for C. glabrata, C. krusei, C. parapsilosis, C. tropicalis; 5,000 CFU/mL for C. albicans and C. auris; 7,500 CFU/mL for E. coli, G. vaginalis, K. pneumoniae and P. mirabilis; 10,000 CFU/mL for E. faecalis; 12,500 CFU/mL for Prevotella spp.; 15,000 CFU/mL for A. baumanii and S. aureus; and 25,000 CFU/mL for B. fragilis and P. aeruginosa (Figure 3B).

Figure 3.
Figure 3.

The analytical sensitivity was assessed in the most frequently found urogenital pathogens (10 bacteria and 6 fungal species) by spiking reference whole organisms into negative urine matrix. (A) Each pathogen was spiked in triplicate per concentration (CFU/mL) followed by genomic DNA extraction. Experiments 1 and 2 included 6 replicates for each concentration of the 10-fold dilution spike ins, ranging from 10 to 1 million CFU/mL. Experiments 3 & 4 were used to fine tune the LoD by testing 6 replicates per concentration, ranging from 5x103 to 5x104 CFU/mL. (B) Summary of replicates detected per concentration tested for each urogenital pathogen. The LoD was reproducibly verified and determined based on a 100% positivity rate resulting in an overall LoD of <25,000 CFU/mL for bacteria and <5,000 CFU/mL for fungi. Specifically, the LoDs determined for each species were 1,000 CFU/mL for C. glabrata, C. krusei, C. parapsilosis and C. tropicalis; 5,000 CFU/mL for C. albicans and C. auris; 7,500 CFU/mL for E. coli, G. vaginalis, K. pneumoniae and P. mirabilis; 10,000 CFU/mL for E. faecalis.; 12,500 CFU/mL for Prevotella spp.;15,000 CFU/mL for A. baumanii and S. aureus; 25,000 CFU/mL for P. aeruginosa and B.fragilis.

REPRODUCIBILITY

Qualitative precision was established with inter- and intra-assay reproducibility experiments in which 19 different contrived pools containing at least 3-6 different microbial species were assessed. To account for genetic variation, a total of 66 strains belonging to 15 species of bacteria and fungi were processed. Each pool was tested in triplicate for three consecutive days obtaining a sensitivity, specificity, and qualitative reproducibility of 100%. Detailed results are presented in Supplemental Table 2.

IN SILICO

The in silico studies assessed BIOTIA-DX’s ability to identify a large, diverse set of microorganisms absent in clinical or contrived samples and determined whether the pipeline would misidentify closely related organisms (i.e., “cross-reactivity”). The sensitivity and specificity obtained is 99.99% (a detailed breakdown of results is summarized in Supplemental Table 3). Only 1.7% of synthetic samples tested (n=147) showed cross-reactivity and most species observed were within the same genus, do not have clinical relevance, or are unlikely to be found in a urine specimen (Supplemental Table 4). Cross-reactivities within the same genus were further differentiated by targeting specific genes containing abundant discriminating sites between species. Cross-reactive samples from 49 species were assigned the correct species in 94% of cases, further reducing false positives to 0.3% (n=27) (Supplemental Table 4). BIOTIA- DX overall performance on the training set, specificity, and in silico studies is shown on Supplemental Figure 2.

Supplemental Figure 1.
Supplemental Figure 1.

Genomic DNA from urine specimens with apparent false positive and false negative results where BIOTIA-ID differed from the original culture result (49.2% of analytes detected), were tested by qPCR or Sanger sequencing (n=123). 87% of NGS results were ultimately found to be correct by the orthogonal adjudication. Bar chart shows comparator adjudicated results for the urogenital pathogens tested with comparator assays. TP: true positive (dark pink), FP: false positive (dark blue), FN: false negative (gray), TN: true negative (light pink).

Supplemental Figure 2.
Supplemental Figure 2.

BIOTIA-DX ML Performance in samples processed on the training (dark blue), in silico (dark pink), accuracy (gray), reproducibility (light pink) and specificity (light blue) studies.

ACCURACY

The accuracy verification study evaluated the precision of the assay in detecting uropathogens in clinical urine specimens compared to the current gold standard diagnostic (culture). Deidentified clinically tested urine specimens were collected from a reference laboratory based on the pathogens diagnosed by culture. A total of 335 samples (167 clinical and 168 contrived) were tested including 35 culture-negative specimens. Culture conducted by the reference laboratories reported the following CFU/mL distributions in the specimens processed: 42.5% with >100,000 CFU/mL, 11.5% with 50-100k CFU/mL, 22.4% with 10-50k CFU/mL, and 23.6% with insignificant growth or no growth.

BIOTIA-ID detected the spiked microorganisms in contrived specimens and predicted a total of 250 urogenital pathogens in the clinical specimens, classifying 83.2% (n=139) as positive for UTI (Figure 4A, B). Single organism infections were observed in 43.1% of the specimens while 40.1% were coinfections and polymicrobial infections (3+ pathogens, 19.8%). Conversely, 72.5% of specimens were considered single organism infections by culture (data not shown). As depicted in Figure 4C, BIOTIA-ID detected 27 different urogenital pathogen species, and as expected due to collection bias, higher frequencies of E. coli, E. faecalis, K. pneumoniae, P. mirabilis and S. aureus were observed. Of these, 22 additional species not reported by culture were predicted with BIOTIA-ID, representing 52.4% (n=131) of the total urogenital pathogens detected by NGS (n=250). Approximately one third (n=39) of these belonged to the 5 key bacterial species used throughout this validation, a further 29.8% (n=39) were uropathogens commonly reportable by culture (e.g., S. agalactiae, Staphylococcus spp. and other gram- negative Enterobacterales species), while 24.4% (n=32) were organisms not usually culturable with the SOC (e.g., Aerococcus spp., G. vaginalis and Prevotella spp.). NGS detected Candida species (C. albicans, C. glabrata, C. tropicalis and C. krusei) in 16% (n=21) of the unreported taxa, while culture reported yeast in only 2 specimens.

Figure 4.
Figure 4.

The BIOTIA-ID Urine NGS Assay accurately detects and classifies urogenital pathogens in clinical (n=167) and contrived (n=168) specimens in the accuracy study. (A) Heatmap depicts the relative abundance of taxa detected by NGS and is annotated with culture diagnosis, and agreement (green) or disagreement (purple) between BIOTIA-DX ML and culture. (B) Pie chart showing the distribution of the number of urogenital pathogens detected in clinical urine specimens tested with BIOTIA-ID. (C) Pie chart depicts the distribution of microbial analytes detected with culture (navy blue) and with NGS (pink). (D) Venn diagram illustrates the distribution and overlap of the analytes tested and detected with comparator testing for adjudicating the discrepancies found between NGS and SOC. Apparent false positive and false negative results from 19 analytes, where BIOTIA-ID differed from the original culture result (50.2% of clinical samples), were tested by qPCR or Sanger sequencing (n=123), and (E) 87% of these NGS results were found to be correct (n=107) as shown on pie chart showing the distribution of adjudicated results. TP: true positive (n=104, dark pink), TN: true negative (n=3, light pink), FP: false positive (n=12, dark blue), FN: false negative (n=4, dark gray).

Of the clinical specimens tested, 50.9% (n=85) revealed a disagreement between NGS and the gold standard and 11.3% (n=19) completely disagreed on the species detected. Due to culture limitations and the high sensitivity of NGS, comparator testing of 123 analytes (qPCR n=82, Sanger sequencing n=39) was conducted to reconcile discrepancies between findings from both methods. A subset of 19 concordant analytes were used as clinical positive controls and 100% of the analytes detected agreed across the three methods (Figure 4D, Supplemental Table 5). Apparent false positives and false negatives from 19 analytes, where BIOTIA-ID differed from the SOC, were tested and 87% of these NGS results were ultimately correct according to the comparator (Figure 4C, D). E. faecalis, K. pneumoniae, E. coli, Candida spp., Aerococcus spp., and G. vaginalis represented the highest frequency of disagreements (65%) between the SOC and NGS. Comparator reconciled results by taxa tested are shown in Supplemental Figure 1. Moreover, analytes adjudicated as false positives by NGS (n=12, 9.7%) mostly belong to G. vaginalis and Anginosus group streptococci, two common opportunistic pathogens of the urogenital tract. However, false negative (n=4, 3.3%) samples with analytes E. faecalis, K. pneumoniae and E coli yielded below threshold BIOTIA-ID prediction probabilities, demonstrating BIOTIA-DX’s low statistical confidence in defining the presence of the pathogen reported by the SOC.

Following comparator adjudication, the accuracy study revealed a 98.4% sensitivity and 99.7% specificity while performance in clinical specimens achieved a 97.2% sensitivity and 99.6% specificity (Table 2). Overall, BIOTIA-ID accomplished >99.9% sensitivity and specificity across all samples and analytes processed, exceeding ∼14.5k analytes evaluated (Table 2). Contrived specimens yielded >99.9% sensitivity and specificity and analyte breakdown are presented in Supplemental Table 6.

View this table:
Table 2. BIOTIA-ID clinical and analytical validation performance characteristics summary by analyte category (Bacteriology, Mycology, Overall).

DISCUSSION

In this study, we developed and validated a clinical-grade NGS assay, BIOTIA-ID, for accurate pathogen identification in urine specimens using ML-classification. BIOTIA-ID overcomes challenges specific to urine molecular diagnostics including sample collection and transport, low gDNA yields, presence of inhibitors, high human background, and taxonomic bias due to suboptimal extraction (47–56). The laboratory process is compatible with UTT, the preservative of choice for standard urine cultures in a clinical setting. DNA extraction was optimized for efficient lysis of gram-positive bacteria and fungi ensuring sufficient gDNA yields from 2 mL of urine, specifically from low biomass samples, such as those collected from males, from catheterized patients, or from patients with prior antibiotic use. Three external controls and one internal control were implemented to validate the process (PC), monitor cross contamination coming from the process or reagents (NEC, NTC), and ensure integrity and validity of each sample minimizing false negatives due to inhibition (IPC). BIOTIA-DX has a clinically curated pathogen database, and the ML-classifier was trained with a combination of urine clinical specimens, clinical and reference isolates, and synthetic metagenomic specimens. Comparator validation of discrepant results along with a comprehensive training set increased the specificity and robustness of pathogen prediction. Analytical performance was assessed on 20 prevalent urogenital pathogens in 1,305 contrived samples, 12,250+ in silico simulations, and 167 clinical specimens. BIOTIA-ID demonstrated high sensitivity (99.92%) and specificity (99.95%) in predicting uropathogens, making it a promising alternative testing to culture-based diagnostics.

Although urine culture is highly reliable with reported ∼90% sensitivity and 86% specificity in healthy outpatient women, its specificity is exceptionally variable in chronically ill patients (76- 95%) and in patients with indwelling catheters where it drops close to 0%(57). BIOTIA-ID’s ability to precisely detect pathogens at low concentrations (LoD <1,000-25,000 CFU/mL) enhances its clinical utility, particularly in the context of managing high-risk immunocompromised patients, rUTI and cUTIs, and/or catheter-associated UTIs (CAUTI). UTI diagnostics currently rely on culture, and while effective, 20-30% of cases go undiagnosed (58–61), with this number increasing in complicated cases. Culture-based diagnostics fail due to numerous issues, including prior antibiotic exposure, collection or storage errors, cross- contamination, selectivity of culture conditions, and inadequate identification of fastidious, anaerobic, and fungal infections (19, 22), many of which can be addressed with NGS-based diagnostics.

In this study, half of the NGS tested clinical specimens revealed a disagreement with culture findings, with most cases detecting more than one urogenital pathogen, including the species reported by culture. About 11.3% of specimens had different species detected with NGS and about 15% exhibited clinically relevant discrepancies like detecting a gram-positive (E. faecalis) instead of a gram-negative pathogen (E. coli). Despite being highly abundant by both NGS and qPCR testing, E. faecalis was the top missed organism by culture. Notably, close to 43% of the clinical specimens which culture diagnosed as P. mirabilis (n=9/21) showed E. faecalis as the dominant pathogen with an NGS-based approach. E. faecalis is frequently under-reported by standard culture with 50% detection rates when compared to Enhanced Quantitative Urine Culture (20, 62). Further, Candida species were missed (n=17) or only reported by culture as yeast (n=2), representing 11.9% of the clinical specimens tested. While Candida can be isolated with the SOC, sensitivity is rather poor. A study compared Candida detection in urine samples with confirmed yeast presence on urinalysis, finding that standard urine culture identified 37% of Candida infections whereas using a fungal-specific media (Sabouraud Dextrose Agar) increased detection to 98%(63). Studies have shown that standard urine culture does not provide optimal growth conditions for all urogenital pathogens causing organisms such as Enterobacterales to outcompete other microbes like Enterococcus and Candida (20, 26). Failure to detect the correct pathogen results in physicians prescribing inadequate antibiotics, thus exacerbating the emergence of antimicrobial resistance, and increasing patient suffering and medical burden(64–68). For example, Enterococcus is the second leading cause of CAUTI, accounting for 45% of CAUTI infections and the third leading cause of hospital acquired UTIs (69–72). Although UTIs represent the primary cause for antibiotic prescriptions in people living in nursing homes, 40 to 75% of cases receive inappropriate antimicrobial therapy (1, 73–75). NGS-based metagenomic tools offer an opportunity for precise infectious disease diagnostics and have the potential to enhance antimicrobial stewardship efforts.

We preferentially procured specimens diagnosed with specific urogenital pathogens that are easily and more consistently cultured with SOC, such as E. coli which accounts for up to 75% of uncomplicated UTIs and 65% of cUTI cases (76–80). Standard urine culture is selective for typical microorganisms that are most commonly responsible for acute and uncomplicated UTIs; nonetheless, augmented culture methods and diagnostic tests are better suited when there is a high clinical suspicion for UTI caused by atypical organisms, as seen in neutropenia (e.g., candiduria), genitourinary tuberculosis, and urinary tract abnormalities (e.g., anaerobic bacteria) (81). Like other molecular-based studies, BIOTIA-ID showed increased detection of atypical pathogens including Candida, G. vaginalis, Aerococcus spp., anaerobic bacteria (Prevotella spp.) and fastidious bacteria, organisms not reported with culture-based diagnostics (30, 82–85). Although the rise of molecular based diagnostics and clinical studies using these technologies have revealed that gram-positive, atypical, or fastidious pathogens account for a larger percentage of cUTI infections, more interventional, and clinical utilization studies are needed to understand the pathogenesis and epidemiology of these understudied microbes (19, 62, 86–88). Despite specimen collection bias, BIOTIA-ID identified a significant number of atypical pathogens in the tested clinical specimens leading us to hypothesize that the detection frequency would likely rise in clinical studies that focus on culture-negative cUTI cases, a population which could greatly benefit from improved diagnostics.

Polymicrobial infections with more than two pathogens are usually considered contamination under the current SOC and reporting of these results as infection is a topic of debate (18). In accordance with other UTI studies employing molecular approaches, about 40 % of clinical specimens tested in this validation showed more than one pathogen detected (83, 85, 89–92). Most clinical algorithms designed for uropathogen detection provide optimal growth conditions for a limited number of microbes and are based on a threshold of 105 CFU/mL. Understandably, many guidelines aim to prevent over prescription of antibiotics. Nonetheless, the development of improved diagnostics can help better guide antimicrobial stewardship efforts while still providing a comprehensive report. It has been demonstrated that presence of certain pathogens, even at low levels, can heighten the virulence of other dominant pathogens and could impact response to antimicrobial therapy (93–99).

Regarding concerns about the potential oversensitivity of an NGS-based assay, BIOTIA-DX was intentionally designed and trained to increase stringency and reduce false positive detection of urogenital commensals or opportunistic pathogens present at colonization levels often a common weakness of many molecular diagnostic tools and clinical studies (34, 81, 88, 100, 101). Relative abundance of microbial species, percentage of human reads, IPC detection and complexity of the microbial profiles are features implemented to enhance the BIOTIA-DX’s ability to correctly predict an infection. This study highlights how ML coupled with extensive laboratory and bioinformatic validation could improve infectious disease diagnostics by addressing the failures of culture-based diagnostics. It is worth noting that we observed a strong linear correlation between spiked CFU/mL and relative abundance of LoD contrived samples (data not shown). The complexity of a clinical urine specimen requires the consideration and development of additional pipeline features for the implementation of a quantification metric that could be relevant and useful to physicians. Additional clinical studies are required to generate a robust dataset, facilitating the creation of an updated, user-friendly NGS-based metric that aligns more closely with current clinical and antibiotic stewardship guidelines for UTI treatment and management (17, 28, 85, 102) Further studies and real-world evaluations are warranted to validate the clinical impact and cost- effectiveness of BIOTIA-ID as a UTI diagnostic tool. The observed performance characteristics suggest potential clinical applications where the advantages of sensitivity, specificity, and comprehensive testing outweigh some of the limitations associated with sequencing cost and TAT. Some use cases that would most benefit from this approach include symptomatic patients with negative culture, cUTI and rUTI, and immunocompromised patients with nonspecific infection symptoms who are at high-risk of developing sepsis.

BIOTIA-ID represents a significant advancement in UTI diagnostics, offering notable advantages over culture-based methods. Its high accuracy, low LoD, and potential to guide targeted antimicrobial therapy support the goals of antimicrobial stewardship. Implementation of BIOTIA-ID in clinical settings has the potential to improve patient outcomes, reduce the misuse of antibiotics, and contribute to the global fight against antimicrobial resistance.

Data Availability

All data produced in the present study are available upon reasonable request to the authors.

ACKNOWLEDGEMENTS

We thank Zaineb Bello for clinical specimen collection and laboratory operation support; Tiara Rivera and Tamara Goncharuk for sequencing support; Yehudah Gruenstein and LabQ Team for clinical specimen collection; Adam Nagyhazy-Horvath, Eszter Szollosy, Patrik Blik, the Abesse Team, Pierre Davidoff, Cory Mason for IT support; Taylor Paisie for bioinformatic support; Talisa Vega-Kline for logistical support.

The BIOTIA-DX software pipeline work used Cromwell On Azure; the Microsoft Genomics supported implementation of the Broad Institute’s Cromwell workflow engine on Azure. The development and validation work used the Extreme Science and Engineering Discovery Environment (XSEDE), which is supported by National Science Foundation grant (ACI- 1548562). Specifically, it used the Bridges system, which is supported by NSF award (ACI- 1445606), at the Pittsburgh Supercomputing Center (PSC).

REFERENCES

  1. 1.↵
    Yang X, Chen H, Zheng Y, Qu S, Wang H, Yi F. 2022. Disease burden and long-term trends of urinary tract infections: A worldwide report.
  2. 2.↵
    Zeng Z, Zhan J, Zhang K, Chen H, Cheng S. 2022. Global, regional, and national burden of urinary tract infections from 1990 to 2019: an analysis of the global burden of disease study 2019. World J Urol 40:755–763.
    OpenUrlCrossRef
  3. 3.↵
    Brumwell A, Sutton G, Lantos PM, Hoffman K, Ruffin F, Brinkac L, Clarke TH, Adams MD, Fowler VG, Fouts DE, Thaden JT. 2023. Escherichia coli ST131 Associated with Increased Mortality in Bloodstream Infections from Urinary Tract Source. J Clin Microbiol doi:10.1128/jcm.00199-23.
    OpenUrlCrossRef
  4. 4.
    Buonaiuto VA, Marquez I, De Toro I, Joya C, Ruiz-Mesa JD, Seara R, Plata A, Sobrino B, Palop B, Colmenero JD. 2014. Clinical and epidemiological features and prognosis of complicated pyelonephritis: A prospective observational single hospital-based study. BMC Infect Dis 14.
  5. 5.
    Ryan J, Neill EO, McLornan L. 2021. Urosepsis and the urologist! Curr Urol 15:39–44.
    OpenUrl
  6. 6.
    Tocut M, Zohar I, Schwartz O, Yossepowitch O, Maor Y. 2022. Short- and long-term mortality in patients with urosepsis caused by Escherichia coli susceptible and resistant to 3rd generation cephalosporins. BMC Infect Dis 22.
  7. 7.↵
    Wagenlehner FME, Pilatz A, Weidner W, Naber KG. 2015. Urosepsis: Overview of the Diagnostic and Treatment Challenges. Microbiol Spectr 3.
  8. 8.↵
    Yamamichi F, Shigemura K, Kitagawa K, Takaba K, Tokimatsu I, Arakawa S, Fujisawa M. 2017. Shock due to urosepsis: A multicentre study. Canadian Urological Association Journal 11:E105–E109.
    OpenUrl
  9. 9.↵
    Bjerklund Johansen TE, Cek M, Naber K, Stratchounski L, Svendsen M V., Tenke P. 2007. Prevalence of Hospital-Acquired Urinary Tract Infections in Urology Departments. Eur Urol 51:1100–1112.
    OpenUrlPubMed
  10. 10.↵
    Maskell RM. 2010. The natural history of urinary tract infection in women. Med Hypotheses 74:802–806.
    OpenUrlCrossRefPubMed
  11. 11.
    Brubaker L, Chai TC, Horsley H, Khasriya R, Moreland RB, Wolfe AJ. 2023. Tarnished gold—the “standard” urine culture: reassessing the characteristics of a criterion standard for detecting urinary microbes. Frontiers in Urology 3.
  12. 12.↵
    Perez-Carrasco V, Soriano-Lerma A, Soriano M, Gutiérrez-Fernández J, Garcia-Salcedo JA. 2021. Urinary Microbiome: Yin and Yang of the Urinary Tract. Front Cell Infect Microbiol. Frontiers Media S.A. doi:10.3389/fcimb.2021.617002.
    OpenUrlCrossRef
  13. 13.
    Wolfe AJ, Brubaker L. 2015. Sterile Urine and the Presence of Bacteria. Eur Urol. Elsevier B.V. doi:10.1016/j.eururo.2015.02.041.
    OpenUrlCrossRefPubMed
  14. 14.↵
    Brubaker L, Wolfe AJ. 2015. The new world of the urinary microbiota in women. Am J Obstet Gynecol 213:644–649.
    OpenUrlCrossRefPubMed
  15. 15.↵
    Lough ME, Shradar E, Hsieh C, Hedlin H. 2019. Contamination in Adult Midstream Clean-Catch Urine Cultures in the Emergency Department: A Randomized Controlled Trial. J Emerg Nurs 45:488–501.
    OpenUrl
  16. 16.
    Polin MR, Kawasaki A, Amundsen CL, Weidner AC, Siddiqui NY. 2017. Do Mixed-Flora Preoperative Urine Cultures Matter? South Med J 110:426–429.
    OpenUrl
  17. 17.↵
    LaRocco MT, Franek J, Leibach EK, Weissfeld AS, Kraft CS, Sautter RL, Baselski V, Rodahl D, Peterson EJ, Cornish NE. 2016. Effectiveness of Preanalytic Practices on Contamination and Diagnostic Accuracy of Urine Cultures: a Laboratory Medicine Best Practices Systematic Review and Meta-analysis. Clin Microbiol Rev 29:105–147.
    OpenUrlCrossRef
  18. 18.↵
    Bilsen MP, Jongeneel RMH, Schneeberger C, Platteel TN, van Nieuwkoop C, Mody L, Caterino JM, Geerlings SE, Köves B, Wagenlehner F, Conroy SP, Visser LG, Lambregts MMC. 2023. Definitions of urinary tract infection in current research: a systematic review. Open Forum Infect Dis doi:10.1093/ofid/ofad332.
    OpenUrlCrossRef
  19. 19.↵
    Franz M, Hörl WH. 1999. Common errors in diagnosis and management of urinary tract infection. I: Pathophysiology and diagnostic techniques. Nephrology Dialysis Transplantation 14:2746–2753.
    OpenUrlCrossRefPubMedWeb of Science
  20. 20.↵
    Price TK, Dune T, Hilt EE, Thomas-White KJ, Kliethermes S, Brincat C, Brubaker L, Wolfe AJ, Mueller ER, Schreckenberger PC. 2016. The Clinical Urine Culture: Enhanced Techniques Improve Detection of Clinically Relevant Microorganisms. J Clin Microbiol 54:1216–1222.
    OpenUrlAbstract/FREE Full Text
  21. 21.
    Nienhouse V, Gao X, Dong Q, Nelson DE, Toh E, McKinley K, Schreckenberger P, Shibata N, Fok CS, Mueller ER, Brubaker L, Wolfe AJ, Radek KA. 2014. Interplay between Bladder Microbiota and Urinary Antimicrobial Peptides: Mechanisms for Human Urinary Tract Infection Risk and Symptom Severity. PLoS One 9:e114185.
    OpenUrlCrossRefPubMed
  22. 22.↵
    Ishihara T, Watanabe N, Inoue S, Aoki H, Tsuji T, Yamamoto B, Yanagi H, Oki M, Kryukov K, Nakagawa S, Inokuchi S, Ozawa H, Imanishi T. 2020. Usefulness of next- generation DNA sequencing for the diagnosis of urinary tract infection. Drug Discov Ther 14:42–49.
    OpenUrl
  23. 23.↵
    Heytens S, De Sutter A, Coorevits L, Cools P, Boelens J, Van Simaey L, Christiaens T, Vaneechoutte M, Claeys G. 2017. Women with symptoms of a urinary tract infection but a negative urine culture: PCR-based quantification of Escherichia coli suggests infection in most cases. Clinical Microbiology and Infection 23:647–652.
    OpenUrlPubMed
  24. 24.↵
    ARC. 2022. Global burden of bacterial antimicrobial resistance in 2019: a systematic analysis. The Lancet 399:629–655.
    OpenUrl
  25. 25.↵
    Klausing BT, Tillman SD, Wright PW, Talbot TR. 2016. The influence of contaminated urine cultures in inpatient and emergency department settings. Am J Infect Control 44:1166–1167.
    OpenUrl
  26. 26.↵
    Hilt EE, McKinley K, Pearce MM, Rosenfeld AB, Zilliox MJ, Mueller ER, Brubaker L, Gai X, Wolfe AJ, Schreckenberger PC. 2014. Urine is not sterile: Use of enhanced urine culture techniques to detect resident bacterial flora in the adult female bladder. J Clin Microbiol 52:871–876.
    OpenUrlAbstract/FREE Full Text
  27. 27.
    Fourcade C, Canini L, Lavigne J-P, Sotto A. 2015. A comparison of monomicrobial versus polymicrobial Enterococcus faecalis bacteriuria in a French University Hospital. European Journal of Clinical Microbiology & Infectious Diseases 34:1667–1673.
    OpenUrlCrossRef
  28. 28.↵
    Smelov V, Naber K, Bjerklund Johansen TE. 2016. Improved Classification of Urinary Tract Infection: Future Considerations. European Urology Supplements 15:71–80.
    OpenUrl
  29. 29.
    Magistro G, Stief CG. 2019. The Urinary Tract Microbiome: The Answer to All Our Open Questions? Eur Urol Focus 5:36–38.
    OpenUrl
  30. 30.↵
    Szlachta-McGinn A, Douglass KM, Chung UYR, Jackson NJ, Nickel JC, Ackerman AL. 2022. Molecular Diagnostic Methods Versus Conventional Urine Culture for Diagnosis and Treatment of Urinary Tract Infection: A Systematic Review and Meta-analysis. Eur Urol Open Sci 44:113–124.
    OpenUrl
  31. 31.
    Schmidt K, Mwaigwisya S, Crossman LC, Doumith M, Munroe D, Pires C, Khan AM, Woodford N, Saunders NJ, Wain J, O’Grady J, Livermore DM. 2017. Identification of bacterial pathogens and antimicrobial resistance directly from clinical urines by nanopore- based metagenomic sequencing. Journal of Antimicrobial Chemotherapy 72:104–114.
    OpenUrlCrossRefPubMed
  32. 32.↵
    Lagier J-C, Dubourg G, Million M, Cadoret F, Bilen M, Fenollar F, Levasseur A, Rolain J- M, Fournier P-E, Raoult D. 2018. Culturing the human microbiota and culturomics. Nat Rev Microbiol 16:540–550.
    OpenUrlCrossRefPubMed
  33. 33.↵
    Chen J, Sun L, Liu X, Yu Q, Qin K, Cao X, Gu J. 2022. Metagenomic Assessment of the Pathogenic Risk of Microorganisms in Sputum of Postoperative Patients With Pulmonary Infection. Front Cell Infect Microbiol 12:855839.
    OpenUrl
  34. 34.↵
    Yoo J-J, Shin HB, Song JS, Kim M, Yun J, Kim Z, Lee YM, Lee SW, Lee KW, Kim W Bin, Ryu CB, Park S-W, Park SK, Song H-Y, Kim YH, On The Behalf Of Sms Soonchunhyang Microbiome Multi-Disciplinary Study Group. 2021. Urinary Microbiome Characteristics in Female Patients with Acute Uncomplicated Cystitis and Recurrent Cystitis. J Clin Med 10.
  35. 35.
    Eberly AR, Beebout CJ, Carmen Tong CM, Van Horn GT, Green HD, Fitzgerald MJ, De S, Apple EK, Schrimpe-Rutledge AC, Codreanu SG, Sherrod SD, McLean JA, Clayton DB, Stratton CW, Schmitz JE, Hadjifrangiskou M. 2020. Defining a Molecular Signature for Uropathogenic versus Urocolonizing Escherichia coli: The Status of the Field and New Clinical Opportunities. J Mol Biol 432:786–804.
    OpenUrlCrossRef
  36. 36.↵
    Minich JJ, Sanders JG, Amir A, Humphrey G, Gilbert JA, Knight R. 2019. Quantifying and Understanding Well-to-Well Contamination in Microbiome Research. mSystems 4.
  37. 37.↵
    Wood DE, Lu J, Langmead B. 2019. Improved metagenomic analysis with Kraken 2. Genome Biol 20.
  38. 38.↵
    Garrido-Sanz L, Senar MÀ, Piñol J. 2022. Drastic reduction of false positive species in samples of insects by intersecting the default output of two popular metagenomic classifiers. PLoS One 17.
  39. 39.
    McIntyre ABR, Ounit R, Afshinnekoo E, Prill RJ, Hénaff E, Alexander N, Minot SS, Danko D, Foox J, Ahsanuddin S, Tighe S, Hasan NA, Subramanian P, Moffat K, Levy S, Lonardi S, Greenfield N, Colwell RR, Rosen GL, Mason CE. 2017. Comprehensive benchmarking and ensemble approaches for metagenomic classifiers. Genome Biol 18.
  40. 40.
    Peabody MA, Van Rossum T, Lo R, Brinkman FSL. 2015. Evaluation of shotgun metagenomics sequence classification methods using in silico and in vitro simulated communities. BMC Bioinformatics 16.
  41. 41.↵
    Ye SH, Siddle KJ, Park DJ, Sabeti PC. 2019. Benchmarking Metagenomics Tools for Taxonomic Classification. Cell. Cell Press doi:10.1016/j.cell.2019.07.010.
    OpenUrlCrossRef
  42. 42.↵
    Chen S, Zhou Y, Chen Y, Gu J. 2018. Fastp: An ultra-fast all-in-one FASTQ preprocessor. Bioinformatics 34:i884–i890.
    OpenUrlCrossRefPubMed
  43. 43.
    Langmead B, Salzberg SL. 2012. Fast gapped-read alignment with Bowtie 2. Nat Methods 9:357–359.
    OpenUrlCrossRefPubMedWeb of Science
  44. 44.
    Li H. 2018. Minimap2: Pairwise alignment for nucleotide sequences. Bioinformatics 34:3094–3100.
    OpenUrlCrossRefPubMed
  45. 45.
    Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, Marth G, Abecasis G, Durbin R. 2009. The Sequence Alignment/Map format and SAMtools. Bioinformatics 25:2078–2079.
    OpenUrlCrossRefPubMedWeb of Science
  46. 46.↵
    Zaharia M, Bolosky WJ, Curtis K, Fox A, Patterson D, Shenker S, Stoica I, Karp RM, Sittler T. 2011. Faster and More Accurate Sequence Alignment with SNAP.
  47. 47.↵
    Karstens L, Asquith M, Caruso V, Rosenbaum JT, Fair DA, Braun J, Gregory WT, Nardos R, McWeeney SK. 2018. Community profiling of the urinary microbiota: considerations for low-biomass samples. Nat Rev Urol 15:735–749.
    OpenUrl
  48. 48.
    Mahony J, Chong S, Jang D, Luinstra K, Faught M, Dalby D, Sellors J, Chernesky M. 1998. Urine Specimens from Pregnant and Nonpregnant Women Inhibitory to Amplification of Chlamydia trachomatis Nucleic Acid by PCR, Ligase Chain Reaction, and Transcription-Mediated Amplification: Identification of Urinary Substances Associated with Inhibition and Removal of Inhibitory Activity. J Clin Microbiol 36:3122–3126.
    OpenUrlAbstract/FREE Full Text
  49. 49.
    Ackerman AL, Anger JT, Khalique MU, Ackerman JE, Tang J, Kim J, Underhill DM, Freeman MR. 2019. Optimization of DNA extraction from human urinary samples for mycobiome community profiling. PLoS One 14:e0210306.
    OpenUrl
  50. 50.
    Khan G, Kangro HO, Coates PJ, Heath RB. 1991. Inhibitory effects of urine on the polymerase chain reaction for cytomegalovirus DNA. J Clin Pathol 44:360–365.
    OpenUrlAbstract/FREE Full Text
  51. 51.
    Munch MM, Chambers LC, Manhart LE, Domogala D, Lopez A, Fredricks DN, Srinivasan, S. 2019. Optimizing bacterial DNA extraction in urine. PLoS One 14:e0222962.
    OpenUrl
  52. 52.
    Nearing JT, Comeau AM, Langille MGI. 2021. Identifying biases and their potential solutions in human microbiome studies. Microbiome 9:113.
    OpenUrlCrossRef
  53. 53.
    McLaren MR, Willis AD, Callahan BJ. 2019. Consistent and correctable bias in metagenomic sequencing experiments. Elife 8.
  54. 54.
    Karstens L, Asquith M, Davin S, Fair D, Gregory WT, Wolfe AJ, Braun J, McWeeney S. 2019. Controlling for Contaminants in Low-Biomass 16S rRNA Gene Sequencing Experiments. mSystems 4.
  55. 55.
    Mahony J, Chong S, Jang D, Luinstra K, Faught M, Dalby D, Sellors J, Chernesky M. 1998. Urine Specimens from Pregnant and Nonpregnant Women Inhibitory to Amplification of Chlamydia trachomatis Nucleic Acid by PCR, Ligase Chain Reaction, and Transcription-Mediated Amplification: Identification of Urinary Substances Associated with Inhibition and Removal of Inhibitory Activity. J Clin Microbiol 36:3122–3126.
    OpenUrlAbstract/FREE Full Text
  56. 56.↵
    Khan G, Kangro HO, Coates PJ, Heath RB. 1991. Inhibitory effects of urine on the polymerase chain reaction for cytomegalovirus DNA. J Clin Pathol 44:360–365.
    OpenUrlAbstract/FREE Full Text
  57. 57.↵
    Chan-Tack KM, Trautner BW, Morgan DJ. 2020. The varying specificity of urine cultures in different populations. Infect Control Hosp Epidemiol 41:489–491.
    OpenUrl
  58. 58.↵
    Heytens S, Boelens J, Claeys G, DeSutter A, Christiaens T. 2017. Uropathogen distribution and antimicrobial susceptibility in uncomplicated cystitis in Belgium, a high antibiotics prescribing country: 20-year surveillance. European Journal of Clinical Microbiology & Infectious Diseases 36:105–113.
    OpenUrlCrossRefPubMed
  59. 59.
    De Backer D, Christiaens T, Heytens S, De Sutter A, Stobberingh EE, Verschraegen G. 2008. Evolution of bacterial susceptibility pattern of Escherichia coli in uncomplicated urinary tract infections in a country with high antibiotic consumption: a comparison of two surveys with a 10 year interval. Journal of Antimicrobial Chemotherapy 62:364–368.
    OpenUrlCrossRefPubMedWeb of Science
  60. 60.
    Heytens S, De Sutter A, Coorevits L, Cools P, Boelens J, Van Simaey L, Christiaens T, Vaneechoutte M, Claeys G. 2017. Women with symptoms of a urinary tract infection but a negative urine culture: PCR-based quantification of Escherichia coli suggests infection in most cases. Clinical Microbiology and Infection 23:647–652.
    OpenUrlPubMed
  61. 61.↵
    Naber KG, Schito G, Botto H, Palou J, Mazzei T. 2008. Surveillance Study in Europe and Brazil on Clinical Aspects and Antimicrobial Resistance Epidemiology in Females with Cystitis (ARESC): Implications for Empiric Therapy. Eur Urol 54:1164–1178.
    OpenUrlCrossRefPubMedWeb of Science
  62. 62.↵
    Hochstedler BR, Burnett L, Price TK, Jung C, Wolfe AJ, Brubaker L. 2022. Urinary microbiota of women with recurrent urinary tract infection: collection and culture methods. Int Urogynecol J 33:563–570.
    OpenUrl
  63. 63.↵
    Helbig S, Achkar JM, Jain N, Wang X, Gialanella P, Levi M, Fries BC. 2013. Diagnosis and inflammatory response of patients with candiduria. Mycoses 56:61–69.
    OpenUrlCrossRefPubMed
  64. 64.↵
    Durkin MJ, Keller M, Butler AM, Kwon JH, Dubberke ER, Miller AC, Polgreen PM, Olsen MA. 2018. An assessment of inappropriate antibiotic use and guideline adherence for uncomplicated urinary tract infections. Open Forum Infect Dis 5.
  65. 65.
    Kaye KS, Gupta V, Mulgirigama A, Joshi A V., Scangarella-Oman NE, Yu K, Ye G, Mitrani-Gold FS. 2021. Antimicrobial Resistance Trends in Urine Escherichia coli Isolates From Adult and Adolescent Females in the United States From 2011 to 2019: Rising ESBL Strains and Impact on Patient Management. Clinical Infectious Diseases 73:1992– 1999.
    OpenUrlCrossRef
  66. 66.
    Moon RC, Marijam A, Mitrani-Gold FS, Gibbons DC, Kartashov A, Rosenthal NA, Joshi A V. 2022. Treatment patterns, healthcare resource use, and costs associated with uncomplicated urinary tract infection among female patients in the United States. PLoS One 17.
  67. 67.
    Shafrin J, Marijam A, Joshi A V., Mitrani-Gold FS, Everson K, Tuly R, Rosenquist P, Gillam M, Ruiz ME. 2022. Impact of suboptimal or inappropriate treatment on healthcare resource use and cost among patients with uncomplicated urinary tract infection: an analysis of integrated delivery network electronic health records. Antimicrob Resist Infect Control 11.
  68. 68.↵
    Zhu H, Chen Y, Hang Y, Luo H, Fang X, Xiao Y, Cao X, Zou S, Hu X, Hu L, Zhong Q. 2021. Impact of inappropriate empirical antibiotic treatment on clinical outcomes of urinary tract infections caused by Escherichia coli: a retrospective cohort study. J Glob Antimicrob Resist 26:148–153.
    OpenUrl
  69. 69.↵
    Bizuayehu H, Bitew A, Abdeta A, Ebrahim S. 2022. Catheter-associated urinary tract infections in adult intensive care units at a selected tertiary hospital, Addis Ababa, Ethiopia. PLoS One 17.
  70. 70.
    Dunny GM, Hancock LE, Shankar N. 2014. Enterococcal Biofilm Structure and Role in Colonization and Disease.
  71. 71.
    Lin E, Bhusal Y, Horwitz D, Shelburne SA, Trautner BW. 2012. Overtreatment of Enterococcal Bacteriuria. Arch Intern Med 172:33.
    OpenUrlCrossRefPubMedWeb of Science
  72. 72.↵
    Maki D. 2001. Engineering out the Risk of Infection with Urinary Catheters. Emerg Infect Dis 7:342–347.
    OpenUrlCrossRefPubMedWeb of Science
  73. 73.↵
    D’Agata ED, Loeb MB, Mitchell SL. 2013. Challenges in assessing nursing home residents with advanced dementia for suspected urinary tract infections. J Am Geriatr Soc 61:62–66.
    OpenUrlCrossRefPubMed
  74. 74.
    Rotjanapan P, Dosa D, Thomas KS. 2011. Potentially Inappropriate Treatment of Urinary Tract Infections in Two Rhode Island Nursing Homes. Arch Intern Med 171.
  75. 75.↵
    Rowe TA, Juthani-Mehta M. 2013. Urinary tract infection in older adults. Aging health doi:10.2217/ahe.13.38.
    OpenUrlCrossRef
  76. 76.↵
    Czaja CA, Scholes D, Hooton TM, Stamm WE. 2007. Population-Based Epidemiologic Analysis of Acute Pyelonephritis. Clinical Infectious Diseases 45:273–280.
    OpenUrlCrossRefPubMedWeb of Science
  77. 77.
    Flores-Mireles AL, Walker JN, Caparon M, Hultgren SJ. 2015. Urinary tract infections: epidemiology, mechanisms of infection and treatment options. Nat Rev Microbiol 13:269– 284.
    OpenUrlCrossRefPubMed
  78. 78.
    Foxman B. 2014. Urinary Tract Infection Syndromes: occurrence, recurrence, bacteriology, risk factors, and disease burden. Infect Dis Clin North Am 28:1–13.
    OpenUrlCrossRefPubMed
  79. 79.
    Olenski S, Scuderi C, Choo A, Bhagat Singh AK, Way M, Jeyaseelan L, John G. 2019. Urinary tract infections in renal transplant recipients at a quaternary care centre in Australia. BMC Nephrol 20.
  80. 80.↵
    Ronald A. 2002. The etiology of urinary tract infection: traditional and emerging pathogens. Am J Med 113:14–19.
    OpenUrlCrossRef
  81. 81.↵
    Xu R, Deebel N, Casals R, Dutta R, Mirzazadeh M. 2021. A new gold rush: A review of current and developing diagnostic tools for urinary tract infections. Diagnostics. MDPI doi:10.3390/diagnostics11030479.
    OpenUrlCrossRef
  82. 82.↵
    Festa RA, Luke N, Mathur M, Parnell L, Wang D, Zhao X, Magallon J, Remedios-Chan M, Nguyen J, Cho T, Ngo A, Murphy M, Baunoch D. 2023. A test combining multiplex- PCR with pooled antibiotic susceptibility testing has high correlation with expanded urine culture for detection of live bacteria in urine samples of suspected UTI patients. Diagn Microbiol Infect Dis 107.
  83. 83.↵
    Jia K, Huang S, Shen C, Li H, Zhang Z, Wang L, Zhao G, Wu Z, Lin Y, Xia H, Tang M, Yang H, Hu H. 2023. Enhancing urinary tract infection diagnosis for negative culture patients with metagenomic next-generation sequencing (mNGS). Front Cell Infect Microbiol 13.
  84. 84.
    Moustafa A, Li W, Singh H, Moncera KJ, Torralba MG, Yu Y, Manuel O, Biggs W, Venter JC, Nelson KE, Pieper R, Telenti A. 2018. Microbial metagenome of urinary tract infection. Sci Rep 8.
  85. 85.↵
    McDonald M, Kameh D, Johnson ME, Johansen TEB, Albala D, Mouraviev V. 2017. A Head-to-Head Comparative Phase II Study of Standard Urine Culture and Sensitivity Versus DNA Next-generation Sequencing Testing for Urinary Tract Infections. Rev Urol 19:213–220.
    OpenUrlCrossRef
  86. 86.↵
    Morand A, Cornu F, Dufour J-C, Tsimaratos M, Lagier J-C, Raoult D. 2019. Human Bacterial Repertoire of the Urinary Tract: a Potential Paradigm Shift. J Clin Microbiol 57.
  87. 87.
    Neugent ML, Hulyalkar N V., Nguyen VH, Zimmern PE, De Nisco NJ. 2020. Advances in Understanding the Human Urinary Microbiome and Its Potential Role in Urinary Tract Infection. mBio 11.
  88. 88.↵
    Yu Y, Sikorski P, Bowman-Gholston C, Cacciabeve N, Nelson KE, Pieper R. 2015. Diagnosing inflammation and infection in the urinary system via proteomics. J Transl Med 13:111.
    OpenUrlCrossRefPubMed
  89. 89.↵
    Almas S, Carpenter RE, Rowan C, Tamrakar VK, Bishop J, Sharma R. 2023. Advantage of precision metagenomics for urinary tract infection diagnostics. Front Cell Infect Microbiol 13.
  90. 90.
    Jaworska MM, Pecyna P, Jaskiewicz K, Rydzanicz M, Kaluzna M, Pawlaczyk K, Ploski R, Nowak-Malczewska DM, Karolak JA, Gajecka M. 2023. Differences in the composition of the bacterial element of the urinary tract microbiome in patients undergoing dialysis and patients after kidney transplantation. Front Microbiol 14.
  91. 91.
    Vollstedt A, Baunoch D, Wojno KJ, Luke N, Cline K. 2020. Multisite Prospective Comparison of Multiplex Polymerase Chain Reaction Testing with Urine Culture for Diagnosis of Urinary Tract Infections in Symptomatic Patients doi:10.29011/JSU-102.100002.
    OpenUrlCrossRef
  92. 92.↵
    Wojno KJ, Baunoch D, Luke N, Opel M, Korman H, Kelly C, Jafri SMA, Keating P, Hazelton D, Hindu S, Makhloouf B, Wenzler D, Sabry M, Burks F, Penaranda M, Smith DE, Korman A, Sirls L. 2020. Multiplex PCR Based Urinary Tract Infection (UTI) Analysis Compared to Traditional Urine Culture in Identifying Significant Pathogens in Symptomatic Patients. Urology 136:119–126.
    OpenUrlPubMed
  93. 93.↵
    Hou J, Wang L, Alm M, Thomsen P, Monsen T, Ramstedt M, Burmølle M. 2022. Enhanced Antibiotic Tolerance of an In Vitro Multispecies Uropathogen Biofilm Model, Useful for Studies of Catheter-Associated Urinary Tract Infections. Microorganisms 10.
  94. 94.
    Orazi G, O’toole GA. 2019. “It Takes a Village”: Mechanisms Underlying Antimicrobial Recalcitrance of Polymicrobial Biofilms.
  95. 95.
    Vollstedt A, Baunoch D, Wolfe A, Luke N, Wojno KJ, Cline K, Belkoff L, Milbank A, Sherman N, Haverkorn R, Gaines N, Yore L, Shore N, Opel M, Korman H, Kelly C, Jafri M, Campbell M, Keating P, Hazelton D, Makhlouf B, Wenzler D, Sabry M, Burks F, Penaranda M, Smith DE, Cacdac P, Sirls L. 2020. Bacterial Interactions as Detected by Pooled Antibiotic Susceptibility Testing (P-AST) in Polymicrobial Urine Specimens.
  96. 96.
    Morales DK, Hogan DA. 2010. Candida albicans Interactions with Bacteria in the Context of Human Health and Disease. PLoS Pathog 6:e1000886.
    OpenUrlCrossRefPubMed
  97. 97.
    Levison ME, Pitsakis PG. 1987. Susceptibility to Experimental Candida albicans Urinary Tract Infection in the Rat. Journal of Infectious Diseases 155:841–846.
    OpenUrlCrossRefPubMedWeb of Science
  98. 98.
    Codelia-Anjum A, Lerner LB, Elterman D, Zorn KC, Bhojani N, Chughtai B. 2023. Enterococcal Urinary Tract Infections: A Review of the Pathogenicity, Epidemiology, and Treatment. Antibiotics 12:778.
    OpenUrl
  99. 99.↵
    Tien BYQ, Goh HMS, Chong KKL, Bhaduri-Tagore S, Holec S, Dress R, Ginhoux F, Ingersoll MA, Williams RBH, Kline KA. 2017. Enterococcus faecalis Promotes Innate Immune Suppression and Polymicrobial Catheter-Associated Urinary Tract Infection. Infect Immun 85.
  100. 100.↵
    Huang R, Yuan Q, Gao J, Liu Y, Jin X, Tang L, Cao Y. 2023. Application of metagenomic next-generation sequencing in the diagnosis of urinary tract infection in patients undergoing cutaneous ureterostomy. Front Cell Infect Microbiol 13.
  101. 101.↵
    Salter SJ, Cox MJ, Turek EM, Calus ST, Cookson WO, Moffatt MF, Turner P, Parkhill J, Loman NJ, Walker AW. 2014. Reagent and laboratory contamination can critically impact sequence-based microbiome analyses. BMC Biol 12:87.
    OpenUrlCrossRefPubMed
  102. 102.↵
    Vandeputte D, Kathagen G, D’hoe K, Vieira-Silva S, Valles-Colomer M, Sabino J, Wang J, Tito RY, De Commer L, Darzi Y, Vermeire S, Falony G, Raes J. 2017. Quantitative microbiome profiling links gut community variation to microbial load. Nature 551:507–511.
    OpenUrlCrossRefPubMed
Back to top
Posted April 07, 2024.
Download PDF

Supplementary Material

Data/Code
Email
Analytical Validation of a Highly Accurate and Reliable Next-Generation Sequencing-Based Urine Assay
Mara Couto-Rodriguez, David C Danko, Heather L Wells, Sol Rey, Xavier Jirau Serrano, John Papciak, P Ford Combs, Gabor Fidler, Christopher E. Mason, Caitlin Otto, Niamh B. O’Hara, Dorottya Nagy-Szakal
medRxiv 2024.04.05.24305286; doi: https://doi.org/10.1101/2024.04.05.24305286
Twitter logo Facebook logo LinkedIn logo Mendeley logo
Citation Tools

Subject Area