Dipstick-based pathogen detection for wastewater surveillance: Variability analysis using gage repeatability and reproducibility

Gage R&R, Wastewater surveillance, Microbial concentration, Paper dipstick, Variance compo- nents I. INTRODUCTION Wastewater surveillance provides real-time insights into spatio-temporal trends in spread of infectious dis- eases and collecting comprehensive epidemiological data for pandemic preparedness and rapid response [1, 2]. Characterized by its heterogeneous composition of bio- logical materials including bodily fluids such as urine, feces, saliva, and respiratory and nasal secretions from asymptomatic and symptomatic individuals, wastewater poses significant challenges for pathogen monitoring [3– 6]. Furthermore, factors such as rainfall, infection preva- lence, and waste discharge contribute to data variability, complicating on-site wastewater surveillance efforts [7–9]. To simplify the workflow of wastewater surveillance, we have developed an instrument-free dipstick method for nucleic acid extraction from microbes in wastewater. Us- ing this method, we have previously demonstrated con- centration and isolation of nucleic acids from different microbes such as SARS-CoV-2, PMMoV, and Phi6 from wastewater samples with recovery efficiency comparable to commercially available kits [10]. This method includes thermal lysis of pathogens, followed by nucleic acid pu- rification using paper dipsticks, leveraging the favorable nucleic acid binding and release kinetics of cellulose ma- ∗ [email protected] † [email protected] ‡ [email protected] trix [11, 12]. However, due to the manual nature of the method, it is important to understand the relative contri- butions of various sources of variation that impact assay variability. Measurement system analysis (MSA) is one method of quantifying variability, that employs various statisti- cal tools essential for examining statistical quality con- trol [13]. It assesses the impact of measurement er- rors on variations in manufacturing processes and evalu- ates the accuracy, precision, and stability of a measure- ment system (MS) [14, 15]. MS errors can be system- atic errors such as bias and non-linearity, or random er- rors associated with repeatability and reproducibility, of- ten assessed through gage repeatability and reproducibil- ity (gage R&R) studies. Gage R&R is a promising tool for assessing the accuracy, precision, and variability in biological assays, especially those involving complex and heterogeneous samples. Unlike traditional statistical

such as ANOVA and t-tests, which primarily fo- cus on comparing means and determining statistical sig- nificance between experimental conditions [16, 17], gage R&R provides a comprehensive analysis of variability by isolating and quantifying the effects of the measurement system, operator handling, and sample characteristics [18]. Thismethodquantifiesrepeatability(intra-operator variation) and reproducibility (inter-operator variation), providing a clear understanding of reliability and ensur- ing consistent performance of the MS across different set- tings. Gage R&R analysis has been recently utilized to study All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted November 8, 2024. ; https://doi.org/10.1101/2024.11.07.24316947doi: medRxiv preprint NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice. 2 variability in a few biological assays. For instance, Betancourt-Rodriguez et al. employed gage R&R for val- idating a novel isothermal microcalorimetry method by analyzing the variability in heat flow measurements per- formed by two different operators [19]. Similarly, gage R&R has been applied to determine the total variation in a new digital technique measuring the volumetric heal- ing process of free gingival grafts surrounding dental im- plants in ten patients [20]. Additionally, in cell engineer- ing, gage R&R method has been used to validate a pro- tocol to express and purifyα-synuclein, with variability analyzed using descriptors from techniques like PAGE, IMA, IEF, HDX, and peptide mapping [21]. Gage R&R has also been applied to study the operator variability for instruments used in biological processes. For instance, gage R&R study has been used to study operator-to- operator variability amidst trained and in-training ana- lysts for assessing the consistency of digital ECG sys- tems and ECG mark placements at group levels [22]. Similarly, commercially available thermocyclers, such as Eco™ Real-Time PCR system (Illumina) have been val- idated by gage R&R studies. [23]. However, most of these gage R&R studies use the AIAG gage R&R ap- proach developed by the Automotive Industry Action Group (AIAG), primarily used in the automotive indus- try and other manufacturing sectors [24]. AIAG gage R&R is suitable for manufacturing environments where the primary concern is the consistency and accuracy of measurement systems used for quality control and pro- cess optimization [25]. Evaluating the measurement pro- cess (EMP) gage R&R is an advanced version of R&R analysis, that incorporates more detailed analysis and modernstatisticalmethodsthatcanbeappliedtoawider range of industries, including biological and environmen- tal studies [26]. EMP gage R&R provides a comprehen- sive analysis of variability, including interactions between different factors [27], which is crucial for understanding how various sources of variability affect the assay perfor- mance in biological contexts. It can accommodate differ- ent measurement designs and conditions, making it more versatile for various biological and environmental condi- tions as compared to other statistical tools [18, 27, 28]. In order to study the repeatability, reproducibility, ac- curacy and precision of the dipstick method to isolate and capture nucleic acids from1 mL of wastewater sam- ple, we have conducted a rigorous multi-operate EMP gage R&R analysis, as illustrated in Figure 1. The key contribution of this work are summarized below: •The dipstick method was analyzed using both a simplified procedure with manual dipstick prepara- tion and RNA elution, and a rigorous method with automated dipstick preparation and more elabo- rate procedure for RNA elution. These assay steps, which could have contributed to variability, were evaluated through EMP gage R&R. The gage R&R analysis for both methods confirmed that the sim- plified dipstick method did not negatively impact performance, validating overall reliability of the method. •The variability of the simplified dipstick method was further evaluated through a multi-operator gage R&R study using wastewater samples from a sewage pumping station at IIT Bombay. The anal- ysisshowedthatthedipstickmethodhasgageR&R values below 30 %, classifying it as a second-class monitor according to EMP guidelines, and indicat- ing that the simplified dipstick method is within acceptable range of variability and reliable for field applications [29]. The utility for field applications was demonstrated by employing the method for de- tecting variations in PMMoV load associated with population density changes due to summer break on campus. FIG. 1. Schematic illustration of multi-operator gage R&R study for variability analysis of dipstick method employed in wastewater surveillance. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted November 8, 2024. ; https://doi.org/10.1101/2024.11.07.24316947doi: medRxiv preprint 3 TABLE I. Wastewater samples used in this study Sample IDa pH Collection window Dipstick

Captured organism Number of operatorsb Purpose 1 7 During break Simplified & Rigorous Phi6 2 (A,B) Comparing simplified and rigorous dipstick method 2 7 Before break Simplified PMMoV, Phi6 3 (A,B,C) Detecting change in PMMoV concentration due to summer break; variability analysis for simplified dipstick method 3 8 4 7 During break5 7 aWastewater samples were collected from the sewage pumping station at IIT Bombay, Mumbai, India. bEach operator performed measurements in triplicates. II. MATERIALS AND METHODS A. Wastewater samples and pre-processing The microbes and the concentration methods used in this study are outlined in Table I. The selected microbes include bacteriophage Phi6, commonly used as a surro- gate for studying enveloped RNA viruses such as SARS- CoV-2 [30], and pepper mild mottle virus (PMMoV), commonly utilized to assess fecal pollution in water sam- ples and microbial water quality. PMMoV is abundantly present in human feces, making it a reliable control or- ganism for normalizing pathogen loads in wastewater samples relative to population size [31]. Wastewater sam- ples were collected following the sampling strategy out- lined by the CDC [32], from the sewage pumping station at the IIT Bombay campus, Mumbai, India, encompass- ing both residential and academic zones. This study utilized5 wastewater samples, denoted by sample IDs 1 to 5. Wastewater sample ID 1 was col- lected to evaluate the performance variation between the simplified and rigorous dipstick method. Sample IDs 2 and 3 were collected before semester break, while sam- ple IDs 4 and 5 were collected during semester break. Sample IDs 2 to 5 were used to study gage R&R using simplified dipstick method and for detecting changes in PMMoV concentrations associated with population fluc- tuations before and during the semester break. For sam- ple IDs 2 to 5, bacteriophage Phi6 served as a process control. A schematic of the sample pre-processing method is shown in supplementary information Figure S1. Briefly, 100 mL of raw grab wastewater samples were aseptically collected using sterile wide-mouth bottles (Tarsons, PP autoclavable). All the samples were collected between 9 am to 10 am in triplicates and stored at4 ◦C before be- ing processed within24 hoursof collection. The wastewa- ter samples were heat inactivated at60 ◦C for 30 min fol- lowed by dechlorination with1 mLof 1 g/10mLof sodium thiosulfate [33, 34]. The pH levels were measured us- ing pH indicator strips (MQuant®, Merck). For the dipstick assay, 1 mL of wastewater samples were spiked with bacteriophage Phi6 to attain a final concentration of 106 PFU/mL (plaque forming units/milliliter). To mimic the natural wastewater conditions, the spiked wastewa- ter samples were gently mixed using a shaker at30 rpm for 10 min at room temperature. The samples were then filtered using a5 µm mixed cellulose ester membrane fil- ter (MF-Millipore) followed by a1.6 µm glass fibre mem- brane filter to eliminate suspended particles. B. Dipstick preparation and method of use Two different dipstick methods, labeled simplified and rigorous, were evaluated with distinct preparation tech- niques and usage protocols. Purification of nucleic acids released in 1 mL of the sample following thermal lysis at 60 ◦C was achieved in three steps as detailed in our previous work [10]: • Step 1 : Nucleic acid binding to the exposed region of the dipstick (i.e., portion of the dipstick that is not coated with wax) • Step 2 : Washing of exposed region of the dipstick to remove impurities • Step 3 : Elution of RNA in reverse transcription master-mix The dipstick preparation was based on the methodol- ogy outlined by Mason & Botella (2020) [35]. For both methods, the dipsticks were prepared using Whatman grade 1 filter paper sheets (100 mm× 75 mm, GE Health- care). Colored candle wax was melted at80 ◦C on a hot- plate until the wax was completely liquefied in a120 mm borosilicate petri dish. The molten wax was infused into the cellulose filter paper, saturating a length up to55 mm from one end to create a waterproof handle for holding the dipstick. Photographs of dipsticks prepared for sim- plified and rigorous methods are shown in supplementary information Figure S2. The subsequent preparation and usage for each method are described below: All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted November 8, 2024. ; https://doi.org/10.1101/2024.11.07.24316947doi: medRxiv preprint 4 •Simplified dipstick assay: For the simplified method, wax-infusedcellulosesheetsweremanually cut using sterile scissors. The dipsticks were cut to a width of 2 mm and an overall length of65 mm, including the wax-infused segment. The exposed section responsible for nucleic acid capture had a length of 10 mm, with the length of the triangular section at the tip being1 mm. Initially, the cellulose dipstick was immersed in 1 mL of heat-lysed sample and dipped 10 to 15 times till it is completely soaked to facilitate nu- cleic acid capture. Subsequently, the dipstick was washed in500 µL of wash buffer (10 mMTris buffer adjusted to pH 8.0) by gently dipping it thrice to remove cellular debris and impurities. Finally, the dipstick was immersed in a0.2 mL tube containing 20 µL reverse transcription master-mix and vigor- ously moved up and down10 to 15 times. •Rigorous dipstick assay: The rigorous dipstick

was developed to be more systematic, re- duce manual handling, and allow for scalability in wastewater monitoring using dipsticks. For this method, wax-infused cellulose sheets were cut into 2 mm wide strips using a pasta maker as orig- inally demonstrated by Mason et.al. [35] The steps for RNA isolation and purification from ther- mally lysed samples were similar to the simplified method, but with a few modifications. Initially, the dipstick was immersed in the sample for45 s followed by washing in 500 µL of TE wash buffer thrice. The RNA captured on the dipstick was then eluted by immersing it for60 s in 40 µL of reverse transcription master-mix . The reverse transcription was performed using (iScript cDNA synthesis kit, Bio-Rad Laboratories) with40 units of RNaseOUT™ recombinant ribonuclease inhibitor (In- vitrogen). The steps involved in cDNA synthesis reaction are detailed in supplementary information Table S1. C. Real-time quantitative PCR (qPCR) for target amplification Toconfirmthepresence andquantifypeppermildmot- tle virus (PMMoV) and bacteriophage Phi6 in wastew- ater samples, real-time quantitative PCR (qPCR) was performed. For benchmarking the dipstick-based RNA extraction methods, RNA was also extracted from the wastewater samples using the Qiagen PowerWa- ter® RNeasy® kit, according to the manufacturer’s in- structions. Briefly, pre-processed samples were filtered through a 0.2 µm nylon filter (Axiva Sichem) using a vacuum filtration assembly, followed by mechanical bead beating for cell lysis. The purified RNA was eluted in 100 µL of RNase-free water, followed by cDNA synthe- sis. The cDNA generated from three methods (Qiagen kit-based RNA extraction, simplified dipstick assay, and rigorous dipstick assay) served as templates for qPCR amplification. No template control (NTC) was included ineachqPCRrun. AllqPCRreactionswereperformedin triplicates. TheqPCRreactionwasperformedusingBril- liant III Ultra-Fast SYBR® Green QPCR Master Mix (Agilent Technologies) in Agilent technologies Stratagene MX3000P for 40 cycles. Details of the primers for PM- MoV and bacteriophage Phi6, as well as the qPCR ther- mal cycling conditions are detailed in Table S2, and S3 in supplementary material. D. Multi-operator study and gage R&R variability analysis To evaluate the accuracy and precision of the dipstick

in pathogen detection from wastewater samples, we conducted a multi-operator study combined with a gage repeatability and reproducibility (Gage R&R) anal- ysis using the EMP (Evaluating the Measurement Pro- cess) approach. While traditional gage R&R studies as- sess repeatability and reproducibility, EMP goes further by focusing on a system’s overall performance and its suitability for reliably measuring a specific process or output [36, 37]. This analysis was performed using the JMP® statistical software, focusing on variability across operators, samples, and interactions. The multi-operator study was conducted to: •Compare the variability between the simplified and rigorous dipstick methods: Two operators (A and B) operated both the dipstick assays in trip- licates on wastewater sample ID 1 spiked with 106 PFU/mL of bacteriophage Phi6 •Assess operator variability and field applicability for the simplified dipstick method: Three operators (A, B, and C) analyzed wastewater samples ID2 to 5 collected before and during the semester break to assess repeatability and reproducibility, and evalu- ate reliability for detecting change in PMMoV load as compared to process control We performed a crossed gage R&R study, wherein each operator performed the dipstick method on each sample in triplicates under similar conditions. This crossed ap- proach completely overlaps each factor, such as operator and sample, allowing for comprehensive variability anal- ysis across these interactions [38, 39]. The input param- eters included part (wastewater sample with unknown concentrations of PMMoV), operator (technicians A, B and C), andmeasurement (threshold cycle (Ct) value ob- tained through RT-qPCR serving as the primary quan- titative measure for gage R&R analysis). Additionally, the sigma multiplier was set to6 standard deviations (±3 sigma), representing the process width. This multiplier is standard in EMP studies to capture a broad range of natural process variation in assay performance [37]. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted November 8, 2024. ; https://doi.org/10.1101/2024.11.07.24316947doi: medRxiv preprint 5 FIG. 2. (a) Comparison of average Ct values for bacteriophage Phi6 RNA captured using the rigorous and simplified dipstick

by operators A and B in sample ID 1. (b) One-way analysis of Ct values for bacteriophage Phi6 across samples 2 to 4, comparing the simplified dipstick method with the Qiagen PowerWater® RNeasy® kit method. Residual Maximum Likelihood (REML) was applied to estimate the variance components, as it effectively man- ages unbalanced data and offers robust variance com- ponent estimates in biological assays [40]. The variance componentsassessedincludedrepeatability, reproducibil- ity, and interaction variation [41]. Repeatability (intra- operator variability) or equipment variation corresponds to thewithin-assay variation in the variance components. A measurement is considered repeatable if a single oper- ator consistently obtains similar results when measuring the same sample multiple times. Reproducibility (inter- operator variation) or appraiser variability is analogous to between-operator variability in the variance compo- nents and measures precision across different operators. If multiple operators achieve consistent results using the same dipstick method on the same sample, the measure- ment is considered reproducible. Additionally, we evalu- ated product variation (or sample variation), which ex- plains the variability due to factors like dipstick manu- facturing, sample heterogeneity, and differences in viral load across samples. Operator-sample interaction or in- teraction variation, which accounts for the variability in- troduced by the operator and sample combination, was also assessed. Initially, operators A and B assessed the variability in simplified and rigorous dipstick method. Once it was established that the variation in both the methods was comparable, a more thorough study was conducted with three operators A, B and C to study the simpli- fied dipstick method to track change in PMMoV load with change in population. The two-operator analysis for sample ID1 included 36 measurements (1 wastewater sample processed in triplicate) with two operators test- ing both dipstick methods, and Ct measurements further obtained in triplicates. Similarly, the gage R&R study for sample IDs2 to 5 to assess operator-to-operator vari- ability involved 108 measurements (4 wastewater sam- ples processed in triplicates), each tested three times by three different operators, and Ct measurements further obtained in triplicates. Details of the samples are pre- sented in Table I. III. RESULTS AND DISCUSSION A. Simplified dipstick method performs at par with the rigorous dipstick method Variabilityanalysiswasperformedfortheisolationand purificationofbacteriophagePhi6spikedin 1 mLwastew- ater (sample ID 1), to attain a final concentration of 106 PFU/mL. As shown in Figure 2(a), the average Ct values for sample ID 1 collected and processed in tripli- cates by operators A and B were consistent with a grand mean of 25.14. The variance components analysis as shown in supplementary information Figure S3 revealed that there is negligible variation due to the method or operator handling. Similarly, the operator-method inter- action was minimal, confirming there was negligible op- erator and method dependency. The total observed vari- ation was primarily due to repeatability. These results agree with the EMP gage R&R analysis shown in sup- plementary information Figure S3, which demonstrated minimal operator-to-operator variability. The product variation (which accounts for variation due to the dip- stick manufacturing) and interaction variation (variation All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted November 8, 2024. ; https://doi.org/10.1101/2024.11.07.24316947doi: medRxiv preprint 6 due to operator-sample combination) were negligible. Following the comparison between the simplified and rigorous dipstick methods, we evaluated the performance of the simplified dipstick method against the Qiagen PowerWater® RNeasy® kit for detecting bacteriophage Phi6 in samples IDs2 to 4. The one-way analysis mean diamond plots in Figure 2(b), indicate that the spread in Ct values of both simplified dipstick and commercial kit based method are comparable. Additionally, Figure 2(b) and the distribution chart in supplementary information Figure S4 show that the mean Ct value for the kit-based

was22.45 with a standard deviation of0.49 while the simplified dipstick method had a mean Ct value of 23.70 with a comparable standard deviation of0.51. B. Simplified dipstick method has acceptable variability to detect changes in viral load in wastewater samples To further study variability sources in the simplified dipstick method, we conducted a multi-operator study to detect variations in PMMoV concentration correspond- ing to population change due to summer break on IIT Bombay campus. As shown in Figure 3, supported by Figure 4(a), the average Ct values for each operator in- creased in samples collected during the semester break (sample IDs 4 and 5) as compared to those collected be- fore the break (sample IDs 2 and 3), indicating a decrease in PMMoV concentration during summer break. Addi- tionally, as shown in Figure 4(a), the mean average of Ct by operator, and the Ct value spread for all operators was consistent (approximately 33) across all the sam- ples. These observations were further corroborated by the EMP gage R&R results shown in Figure 3(a), where reproducibility (operator-to-operator variation) was re- markably low at just0.7 % followed by operator-sample IDinteractionthatcontributedmerely 7.1 %. Thelargest contributor to variability was product variation, which accounted for72.8 %. Repeatability, which reflects intra- operator variation, contributed19.4 %. ThevariancecomponentsanalysisshowninFigure3(c) was in agreement with the EMP gage R&R results, with minimal operator variability contributing only0.7 % and the majority of the variability arising from sample ID (72.8 %) and operator-sample ID interaction variation ac- counting for only 7.1 %. The total gage R&R was thus 20.1 %, classifying it as a second-class monitor according to EMP guidelines, indicating that the simplified dipstick

is within acceptable range of variability and re- liable for field applications [29]. For bacteriophage Phi6 (used as process control), the average Ct values were similar across different samples and operators (approximately 24) as shown in Figure 3(b), Figure 4(b), Similarly, as illustrated in Figure 4(b), both the mean Ct values for each operator and the Ct value spread were consistent across all samples. These observations agree with the EMP gage R&R results as shown in Figure 3(b), where both reproducibility and product variation were minimal, indicating negligible variability due to the operators or the dipstick method itself. However, the interaction variation was relatively higher for Phi6 at23.4 %compared to7.1 %for PMMoV. The primary source of variability for Phi6 detection was repeatability (accounting for76.6 %). The variance com- ponents as shown in Figure 3(d), mirrored the EMP gage R&R results. C. Discussion The simplified dipstick method involves manual han- dling, which could introduce operator-to-operator varia- tion, particularly during the elution step. To determine if such variation is significant, we compared the RT-qPCR Ct values as a readout by using Phi6 RNA isolated by simplified and rigorous dipstick method as a template to assess variance components through EMP gage R&R analysis. The analysis assessed variability due to the dipstick method, operator, and operator-method inter- action, as well as the repeatability and reproducibility of the assays. While comparing the simplified and rigorous dipstick methods, the minimal operator-to-operator vari- ability confirms that both the methods produced consis- tent results across operators. Furthermore, the negligible part-to-part (sample-to-sample) variation was in agree- ment with the experimental conditions, as the spiked bacteriophage Phi6 concentration was the same across all the samples. The gage R&R results confirmed that most of the observed variability was due to repeatability, and the mean average Ct for both the dipstick methods was similar. Both the methods were highly consistent in performance, and the simplified dipstick method per- formed at par with the rigorous dipstick method, with no significant performance differences, yielding reproducible

across operators. The minimal product and in- teraction variation further demonstrated that there was no significant variation in the dipstick method or in the interaction between operators and the dipstick methods. Additionally, we compared the kit-based method with simplified dipstick method. It was observed that even though the mean Ct values differ due to difference in re- covery efficiency, the standard deviations for both meth- ods were similar. This indicates comparable variability in Ct values between the two methods, promising the suitability of simplified dipstick method for field applica- tions. Field applicability of simplified dipstick method was further studied by three operators to detect changes in PMMoV load with change in population. All the three operators A, B and C successfully detected PMMoV con- centration declining (i.e., increase in Ct value) in samples collected during semester break. The minimal spread in Ct values for the three operatorswhiledetectingPMMoV and the process control bacteriophage Phi6 as shown in Figure 4(a) and (b), emphasizes minimal operator-to- All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted November 8, 2024. ; https://doi.org/10.1101/2024.11.07.24316947doi: medRxiv preprint 7 FIG. 3. (a) Average Ct values for sample IDs2 to 5, processed by operators A, B, and C, along with EMP gage R&R results for (a) PMMoV, and (b) bacteriophage Phi6. Parallelism plots illustrating change in average Ct for sample IDs2 to 5 processed using simplified dipstick method by operators A, B, and C for (c) PMMoV, and (d) bacteriophage Phi6. operator variability. For bacteriophage Phi6, the con- sistent Ct spread across the samples was expected, as the spiked concentration of bacteriophage Phi6 concen- tration was the same across all samples. The low vari- abilityinoperator-sampleIDinteractionforPMMoVfur- ther demonstrated minimal influence of specific operator handling on sample variability. However, the interac- tion variation for Phi6 was comparatively higher, largely caused by the handling of operator C as demonstrated in the parallelism plots shown in Figure 3(d), which show a slight deviation in Ct values for sample IDs3 and 5. Re- peatability variation for Phi6 reflects intra-operator vari- ation due to manufacturing variations in dipstick, and the different steps involved in executing the assay includ- ing washing and elution. This suggests room for further optimization of the method and dipstick design. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted November 8, 2024. ; https://doi.org/10.1101/2024.11.07.24316947doi: medRxiv preprint 8 FIG. 4. (a) Gage R&R mean plots for simplified dipstick

to detect PMMoV in sample IDs2 to 5, illustrating variation of Ct with various parameters: (i) Ct vs. sample ID, and (ii) Ct vs. operator. (b) Gage R&R mean plots for simplified dipstick method to detect bacteriophage Phi6 in sample IDs 2 to 5, illustrating variation of Ct with various parameters: (i) Ct vs. sample ID, and (ii) Ct vs. operator. The primary source of variability for PMMoV detec- tion observed in product variation was due to the natu- ral fluctuations in PMMoV concentration due to change in population across the different samples collected over time. This high product variation is expected, given the dynamic nature of wastewater viral loads as population density and activity change. The total gage R&R at less than 30 % indicates acceptable variability [27], demon- strating the assay’s reliability in consistently detecting target pathogens, even with moderate operator and sam- ple variation, suitable for field applications. IV. CONCLUSION AND FUTURE WORK This study demonstrates the utility of gage R&R anal- ysis as a robust method to validate a biological assay when applied to complex and heterogeneous environmen- tal samples such as wastewater. By systematically isolat- ing and quantifying variations due to the measurement system, operator handling, and the samples themselves, gage R&R allows for a detailed assessment of both re- peatability and reproducibility. The gage R&R analysis suggested that the variability was primarily attributed to within-sample repeatability, followed by operator-sample interaction and negligible due to operator-to-operator handling, indicating that the dipstick method consis- tently produced stable results across different operators andsamples. Thishighlightstherobustnessofthesimpli- fied dipstick method. Unlike ANOVA and t-tests, which are primarily used to determine statistical significance between experimental conditions, gage R&R provides a nuanced understanding of how both the measurement system and human factors influence assay performance, making it indispensable for quality control and method validation in biological research. This shows that gage R&R extends beyond just the dipstick method evaluated here. Our work shows that it can be widely applied in lab- oratory assays and field-deployed diagnostic methods to ensure the reproducibility of results across different set- tings, operators, and batches. This is particularly rel- evant when developing products or assays that will be used in decentralized or resource-constrained environ- ments, where manual methods and operator variability may have a larger impact. The acceptable variability in the simplified dipstick method to isolate and purify nucleic acid highlight the field applicability of the test. With up-scaled manufacturing the dipstick method can be integrated with a field-portable RT-qPCR devices, or rapid detection techniques like loop-mediated isothermal amplification (LAMP), and a lateral flow assay to ensure on-site wastewater monitoring, making the setup more accessible in low-resource or remote areas. DATA AVAILABILITY The data that support the findings of this study are available upon reasonable request from the authors. ACKNOWLEDGMENT S.A. acknowledges the Prime Minister’s Research Fel- lows (PMRF) Scheme for supporting her Ph.D. fellow- ship. This work was partially supported by grants from Department of Science & Technology (DST), Ministry of Science & Technology, Government of India [grant no. DST/IC/IC-IMPACTS/2022/P-13] and Wadhwani Re- search Centre for Bioengineering (WRCB) at IIT Bom- bay [grant DO/2022-WRCB002-076]. The authors thank Ms. Rutuja Chalke at IIT Bombay for help with prepar- ing Figure 1 for this manuscript. DECLARATION OF COMPETING INTERESTS The authors declare that they have no known com- peting financial interests or personal relationships that could have appeared to influence the work reported in this paper. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted November 8, 2024. ; https://doi.org/10.1101/2024.11.07.24316947doi: medRxiv preprint 9 [1] W.-y. Ng, W. Thoe, R. Yang, W.-p. Cheung, C.-k. Chen, K.-h.To, K.-m.Pak, H.-w.Leung, W.-k.Lai, T.-k.Wong, et al., The city-wide full-scale interactive application of sewage surveillance programme for assisting real-time covid-19 pandemic control–a case study in hong kong, Science of The Total Environment875, 162661 (2023). [2] K. Mao, K. Zhang, W. Du, W. Ali, X. Feng, and H. Zhang, The potential of wastewater-based epidemiol- ogy as surveillance and early warning of infectious disease outbreaks, Current Opinion in Environmental Science & Health 17, 1 (2020). [3] A. Tiwari, S. Adhikari, D. Kaya, M. A. Islam, B. Malla, S. P. Sherchan, A. I. Al-Mustapha, M. Kumar, S. Ag- garwal, P. Bhattacharya, et al., Monkeypox outbreak: Wastewater and environmental surveillance perspective, Science of the total environment856, 159166 (2023). [4] A. Tiwari, E. Radu, N. Kreuzinger, W. Ahmed, and T. Pitkänen, Key considerations for pathogen surveil- lance in wastewater, Science of The Total Environment , 173862 (2024). [5] X. Li, J. Kulandaivelu, Y. Guo, S. Zhang, J. Shi, J. O’Brien, S. Arora, M. Kumar, S. P. Sherchan, R. Honda,et al., Sars-cov-2 shedding sources in wastewa- ter and implications for wastewater-based epidemiology, Journal of hazardous materials432, 128667 (2022). [6] K. Bibby, A. Bivins, Z. Wu, and D. North, Making waves: plausible lead time for wastewater based epidemiology as an early warning system for covid-19, Water Research 202, 117438 (2021). [7] L. Pillay, I. D. Amoah, S. Kumari, and F. Bux, Po- tential and challenges encountered in the application of wastewater-based epidemiology as an early warning sys- tem for covid-19 infections in south africa, Acs Es&T Water2, 2105 (2022). [8] J. Black, P. Aung, M. Nolan, E. Roney, R. Poon, D. Hen- nessy, N. D. Crosbie, D. Deere, A. R. Jex, N. John,et al., Epidemiological evaluation of sewage surveillance as a tool to detect the presence of covid-19 cases in a low case load setting, Science of The Total Environment786, 147469 (2021). [9] J. Saththasivam, S. S. El-Malah, T. A. Gomez, K. A. Jabbar, R. Remanan, A. K. Krishnankutty, O. Ogun- biyi, K. Rasool, S. Ashhab, S. Rashkeev,et al., Covid-19 (sars-cov-2) outbreak monitoring using wastewater-based epidemiology in qatar, Science of the Total Environment 774, 145608 (2021). [10] S. Ahuja, S. Tallur, and K. Kondabagil, Simultaneous mi- crobial capture and nucleic acid extraction from wastew- ater with minimal pre-processing and high recovery ef- ficiency, Science of The Total Environment918, 170347 (2024). [11] Y. Zou, M. G. Mason, Y. Wang, E. Wee, C. Turni, P. J. Blackall, M. Trau, and J. R. Botella, Nucleic acid pu- rification from plants, animals and microbes in under 30 seconds, PLoS biology15, e2003916 (2017). [12] B. Boese, K. Corbino, and R. Breaker, In vitro selection and characterization of cellulose-binding rna aptamers using isothermal amplification, Nucleosides, Nucleotides and Nucleic Acids27, 949 (2008). [13] R. Peruchi, A. d. Paiva, P. Balestrassi, J. Ferreira, and R. Sawhney, Weighted approach for multivariate analysis of variance in measurement system analysis, Precision Engineering 38, 651 (2014). [14] S. S. Kamble, A. Gunasekaran, A. Ghadge, and R. Raut, A performance measurement system for industry 4.0 en- abled smart manufacturing system in smmes-a review and empirical investigation, International journal of pro- duction economics229, 107853 (2020). [15] M. M. Wu, S.-T. Chan, D. Mazumder, D. Tamborini, K. A. Stephens, B. Deng, P. Farzam, J. Y. Chu, M. A. Franceschini, J. Z. Qu,et al., Improved accuracy of cere- bral blood flow quantification in the presence of systemic physiology cross-talk using multi-layer monte carlo mod- eling, Neurophotonics8, 015001 (2021). [16] W. C. Driscoll, Robustness of the anova and tukey- kramer statistical tests, Computers & Industrial Engi- neering 31, 265 (1996). [17] C. Bertinetto, J. Engel, and J. Jansen, Anova simulta- neous component analysis: A tutorial review, Analytica Chimica Acta: X6, 100061 (2020). [18] D. J. Wheeler, An honest gauge r&r study, Manuscript (2009). [19] J. Betancourt-Rodríguez, V. M. Zamora-Gasga, J. A. Ragazzo-Sánchez, J. A. N. Zapata, and M. Calderón- Santoyo, Astandardized methodfor genus colletotrichum characterization by isothermal microcalorimetry using thermokinetic parameters, Journal of Microbiological

204, 106651 (2023). [20] C. Docampo-Vázquez, T. Gragera-Alia, M. Fernández- Domínguez, Á. Zubizarreta-Macho, and J. M. Aragoneses-Lamas, Novel digital technique for mea- suring the volumetric healing process of free gingival grafts surrounding dental implants, Frontiers in Dental Medicine 5, 1372312 (2024). [21] H. Thakkar, S. Chatterjee, P. Saxena, R. Eerla, S. Wagh, A. Khairnar, and R. P. Shah, Cell-engineered recombi- nant α-synuclein: A gage r&r validated protocol, Journal of Proteome Research23, 16 (2023). [22] R. A. Ross, C. M. Foley, H. M. Jones, and M. A. Osinski, A method for assessing and monitoring consistency of nonclinical ecg analysis, Journal of Pharmacological and Toxicological Methods116, 107189 (2022). [23] Illumina, Eco™ thermal and optical systems deliver high precision (Technical Note: Real-Time PCR), https://www.illumina.com/documents/products/ technotes/technote_eco_uniformity.pdf (2024), accessed: October 2024. [24] C. LLC, Ford motor company, general motors corpora- tion,PotentialFailureModeandEffectAnalysis(FMEA) , 67 (2008). [25] E. Bottani, R. Montanari, A. Volpi, and L. Tebaldi, Sta- tistical process control of assembly lines in manufactur- ing, Journal of Industrial Information Integration 32, 100435 (2023). [26] D. J. Wheeler and R. W. Lyday,Evaluating the measure- ment process (SPC Press, Incorporated, 1989). [27] D. J. Wheeler,EMP III using imperfect data (SPS Press, 2006). [28] D. J. Wheeler, Range based analysis of means (SPC Press, 2003). [29] D. J. Wheeler, EMP III (Evaluating the Measurement Process): Using Imperfect Data (SPCPRESS(Statistical All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted November 8, 2024. ; https://doi.org/10.1101/2024.11.07.24316947doi: medRxiv preprint 10 Process Control), 2006). [30] S. Torii, H. Furumai, and H. Katayama, Applicability of polyethylene glycol precipitation followed by acid guani- dinium thiocyanate-phenol-chloroform extraction for the detection of sars-cov-2 rna from municipal wastewater, Science of The Total Environment756, 143067 (2021). [31] S.-Y. Hsu, M. Bayati, C. Li, H.-Y. Hsieh, A. Belenchia, J. Klutts, S. A. Zemmer, M. Reynolds, E. Semkiw, H.-Y. Johnson, T.Foley, C.G.Wieberg, J.Wenzel, M.C.John- son, and C.-H. Lin, Biomarkers selection for population normalization in sars-cov-2 wastewater-based epidemiol- ogy, Water research223, 118985 (2022). [32] CDC, Developing a wastewater surveillance sampling strategy (Centers for Disease Control and Prevention, 2020). [33] M. Lau, P. Monis, G. Ryan, A. Salveson, N. Fontaine, J. Blackbeard, S. Gray, and P. Sanciolo, Selection of surrogate pathogens and process indicator organisms for pasteurisation of municipal wastewater—a survey of lit- erature data on heat inactivation of pathogens, Process Safety and Environmental Protection133, 301 (2020). [34] K. M. Babler, A. Amirali, M. E. Sharkey, S. L. Williams, M. M. Boone, G. A. Cosculluela, B. B. Currall, G. S. Grills, J. Laine, C. E. Mason,et al., Comparison of elec- tronegative filtration to magnetic bead-based concentra- tion and v2g-qpcr to rt-qpcr for quantifying viral sars- cov-2 rna from wastewater, ACS ES&T Water2, 2004 (2022). [35] M. G. Mason and J. R. Botella, Rapid (30-second), equipment-free purification of nucleic acids using easy- to-make dipsticks, Nature Protocols15, 3663 (2020). [36] B. S. Ramirez and J. G. Ramírez,Douglas Montgomery’s

to Statistical Quality Control: A JMP Com- panion (Sas Institute, 2018). [37] M. A. Durivage, Practical attribute and variable mea- surement systems analysis (MSA): a guide for conducting Gage R&R studies and test method validations (Quality Press, 2015). [38] F. d. Almeida, T. De Paula, R. Leite, G. Gomes, J. Gomes, A. Paiva, and P. Balestrassi, A multivariate gr&r approach to variability evaluation of measuring in- struments in resistance spot welding process, Journal of Manufacturing Processes36, 465 (2018). [39] R. A. M. Marques, R. B. D. Pereira, R. S. Peruchi, L. C. Brandão, J. R. Ferreira, and J. P. Davim, Multivariate gr&r through factor analysis, Measurement151, 107107 (2020). [40] R. Rameez, S. Jahageerdar, J. Jayaraman, T. I. Chanu, R. Bangera, and A. Gilmour, Evaluation of alternative

for estimating the precision of reml-based esti- mates of variance components and heritability, Heredity 128, 197 (2022). [41] R. S. Kenett and G. Shmueli, Clarifying the terminology that describes scientific reproducibility, Nature methods 12, 699 (2015). All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted November 8, 2024. ; https://doi.org/10.1101/2024.11.07.24316947doi: medRxiv preprint