Discrimination of ranges of closely related RNA targets using CRISPR based detection assay developed using machine learning based optimization

Discrimination of ranges of closely related RNA targets using CRISPR based detection assay developed using machine learning based optimization | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Discrimination of ranges of closely related RNA targets using CRISPR based detection assay developed using machine learning based optimization Tomasz A. Leski, Scott N. Dean, Zachary T. Johnson, Christopher M. Green, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7041916/v1 This work is licensed under a CC BY 4.0 License Status: Under Revision Version 1 posted 12 You are reading this latest preprint version Abstract We have developed a CRISPR/Cas based assay able to distinguish between two ranges of closely related RNA targets using two detection channels. This required a pipeline to design RNA guide sets with the right degree of specificity. We tested our approach using SARS-CoV-2 and zoonotic near-neighbor sarbecoviruses. Using pre-existing guide design rules, we utilized a machine learning based model to design and optimize guide sets for specific detection of SARS-CoV-2 and separately to its nearest neighbors. The in vitro testing of the guide sequences has shown that Cas13 assays can tolerate more mismatches than assumed based on previous guide design rules. Mismatches located closer to the 3’ end of the guide and mismatches evenly distributed throughout the guide resulted in a smaller impact on the guide’s ability to activate the Cas enzyme. Modified SHERLOCK assay for detection and discrimination of SARS-CoV-2 and its zoonotic coronaviruses was developed using optimized sets of guides. The final assay was able to classify the targets into three classes 1) SARS-Co-V2, 2) closest known SARS-Co-V2 near-neighbor BANAL-236 and 3) the remaining zoonotic near-neighbors. This approach provides value through early detection of natural and engineered variants. Biological sciences/Biological techniques Biological sciences/Biotechnology Biological sciences/Computational biology and bioinformatics Biological sciences/Microbiology Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Introduction The emergence of infectious pathogens capable of causing global pandemics has accelerated in recent years, driven by several interconnected factors. The expanding human population, particularly in tropical regions, has led to encroachment into natural habitats, increasing interactions between humans and wildlife. This heightened contact elevates the risk of zoonotic diseases 1 , 2 . Several recently discovered viruses such as Hendra, Nipah, SARS-CoV, MERS-CoV and SARS-CoV-2 have been linked to wildlife reservoirs 3 – 5 . The thriving, largely illegal, and often unregulated wildlife trade further exacerbates the risk of zoonotic transmission and introduces new pathogens into human populations 5 – 7 . Beyond natural zoonoses, there is a growing threat from engineered or synthetic pathogens used as biological weapons. Advancements in genetic engineering and synthetic biology have made development of new pathogens more accessible and affordable, raising concerns about the potential misuse of these technologies by state and non-state actors 8 , 9 . The increasing availability of these technologies heightens the risk of engineered pathogens leading to potential "synthetic pandemics" in the coming decade 10 . Early detection and identification of pathogens are crucial for effective strategies to mitigate both natural and man-made pandemic threats 11 . Recent advancements in molecular biology have significantly enhanced the sensitivity and specificity of pathogen detection methods, enabling rapid responses to infectious disease outbreaks 12 . Techniques such as polymerase chain reaction (PCR), various isothermal amplification methods, high-throughput sequencing, CRISPR-based detection, and immuno-detection technologies have been pivotal in this progress. However, a significant limitation of many current detection technologies is their narrow focus on one or a limited number of specific targets. This approach can lead to delays in identifying outbreaks caused by pathogens that are closely related but not identical to the intended targets, including engineered pathogens. CRISPR/Cas based assays, specifically the ones using Cas12 and Cas13 are capable of amplifying DNA or RNA target detection through the activation of their non-specific nuclease capability upon target binding. The SHERLOCK (Specific High-sensitivity Enzymatic Reporter unLOCKing) assay is the most well-known implementation of the Cas13 based assay and was clinically validated for detection of many different pathogens including SARS-CoV-2 13 . The assay readout usually relies on detection of fluorescent signal generated by digestion of RNA based molecular beacons by the activated non-specific nuclease. Both commercial and novel readout methodologies are being developed for improved detection or point of care applications 14 , 15 . As a result of these developments the design of proper guides has become a key focus of the field. While the majority of researchers have focused on guide design for cellular engineering (i.e. knockdowns) 16 , 17 , the work on methods of design of guides for RNA detection and its use in bio-surveillance has been also recently expanding 18 , 19 . The goal of this study was to develop a CRISPR based assay capable of detecting a specific target and simultaneously distinguishing it from a broad range of closely related near-neighbor targets. The schematic representation of our assay with two detection channels is presented in Fig. 1 . To achieve this goal we needed to construct a developmental pipeline capable of efficiently designing appropriate RNA guide sets with the right degree of specificity that would allow for the distinction between a specific target (inclusion target set) and near neighbors (exclusion target set). In this work we used the SARS-CoV-2 virus and a collection of zoonotic sarbecoviruses closely related to SARS-CoV-2 as a model system. We applied a previously developed machine-learning-based model to design two types of guide sets for use in Cas13a-based (modified SHERLOCK) detection assays 20 . Our results were then used to retrain the model and resulted in better guide predictions. One guide set was able to specifically detect SARSCoV-2 without significantly cross-reacting with the near-neighbor species and second designed to detect several different clades of SARS-CoV-2 related zoonotic sarbecoviruses which represent the closest neighbors of SARS-CoV-2. The guide sequences were tested in vitro , and a selected group of the guides was used to develop the final two-channel assay. Methods All experimental procedures described in this manuscript were carried out in accordance with relevant guidelines and regulations including biosafety and chemical safety regulations. All nucleic acid sequences used were obtained from publicly available NCBI/GenBank and GISAID ( https://www.gisaid.org/ ) database collections and no human subject research was conducted during this project. No materials classified as select agents were obtained or used in any experiments described in this manuscript. Design of SARS-CoV-2 and other coronavirus target sequences and primers Coronavirus genomes are known for their mosaic structure due to their high recombination rates. Temmam et al. 21 identified several recombination breakpoints in SARS-CoV-2 genome which separate fragments of genome with different origin. This study focused on the fragment 11 as defined by Temmam, which codes for the C-terminal part of the Spike protein and more precisely the receptor binding domain (RBD). This part of the gene is one the most divergent among closest zoonotic near neighbors of SARS-CoV-2. In addition, SARS-CoV-2 contains a Furin cleavage site (FCS) in the middle of this region, a sequence which is absent from all known close neighbors of SARS-CoV-2. The target sequences chosen for this study included sequence of the ancestral SARS-CoV-2 strain Wuhan-Hu-1 and sequences from twelve near neighbor zoonotic isolates representing all major clades of SARS-CoV-2-related lineage of Sarbecoviruses. The phylogenetic relationship of the fragment 11 sequences for isolates selected for the study is shown in Fig. 2 and the information on these isolates: names, origin, accession numbers, lengths and starting as well as ending positions of the target sequences in the genomes of the coronavirus isolates are listed in Table 1 . Table 1 Origin and group assignment of target sequences Isolate Origin Host Accession# Start position End position Length Group Wuhan-Hu-1 China human NC045512.2 23310 23950 641 Black BANAL-236 Laos Rhinolophus marshalli (bat) EPI_ISL_4302647 23272 23900 629 Dark red RaTG13 China Rhinolophus affinis (bat) EPI_ISL_402131 23307 23935 629 Dark red RShSTT182 Cambodia Rhinolophus shameli (bat) EPI_ISL_852604 23228 23856 629 Dark red MP789 China Manis javanica (pangolin) MT121216.1 23156 23784 629 Dark red Guangxi-P4L China Manis javanica (pangolin) EPI_ISL_410538 23281 23909 629 Dark red RsYN04 China Rhinolophus stheno (bat) EPI_ISL_1699444 23220 23842 623 Light blue BANAL-247 Laos Rhinolophus malayanus (bat) EPI_ISL_4302648 23145 23767 623 Green RacCS203 Thailand Rhinolophus acuminatus (bat) MW251308.1 23189 23811 623 Green RsYN03 China Rhinolophus sinicus (bat) EPI_ISL_1699443 23156 23784 629 Dark blue SL-CoVZC45 China Rhinolophus pusillus (bat) MG772933.1 23227 23855 629 Yellow RpYN06 China Rhinolophus pusillus (bat) EPI_ISL_1699446 23237 23865 629 Yellow Rco319 Japan Rhinolophus comutus (bat) LC556375.1 23150 23781 632 Purple The panel B of this figure utilizes a phylogenetic tree reproduced with publishers permission from part of supplementary figure S2 originally published in Temmam, S. et al. Nature 604, 330–336, doi: 10.1038/s41586-022-04532-4 (2022). The tree was minimally modified to indicate the sequences used as targets in this study and to define the color-coded groupings of related targets. RNA target synthesis Synthetic double stranded DNA fragments used for synthesis of the RNA targets used in this study were ordered from Integrated DNA Technologies Inc. (IDT, Coralville, IA). The PCR primers pairs complementary to the ends of the synthetic fragments were designed for each target. T7 RNA polymerase promoter sequences were added to the 5’ end of the forward primer in each primer pair. The PCR primers were also purchased from IDT. The sequences the synthetic DNA fragments and PCR primers are listed in Supplementary Material: Sequences. The target RNA molecules were produced using HiScribe™ T7 Quick High Yield RNA Synthesis Kit (New England Biolabs, Ipswich, MA) as described in our previous work. 20 The synthetic DNA fragments of all coronavirus sequences described above were amplified, using the PCR primers, by FastStart Taq DNA polymerase kit (Millipore-Sigma, Burlington, MA) according to the manufacturer’s instructions. The forward PCR primers included T7 promotor sequences, which were incorporated into the amplicons, Supplementary Fig. 1. The transcription reactions were set up using 2 µL of unpurified DNA amplicon preparation, 2 µL of T7 RNA polymerase, 10 µL of 2x NTP buffer and 16 µL of nuclease-free ddH2O (30 µL of total reaction volume). The transcription reactions were incubated at 37°C for 2 h after which 5 µL Turbo DNAse (ThermoFisher, Grand Island, NY) and 15 µL of nuclease-free ddH 2 O were added (increasing the total volume to 40 µL) and incubated further 30 minutes at 37°C to remove the template DNA. The obtained transcript preparations were cleaned up using RNA Clean and Concentrator 25 kit (Zymo Research, Irvine, CA USA) according to the manufacturer instructions. The RNA concentration was determined using Qubit fluorometer and RNA BR (broad range) assay kit (ThermoFisher). The template solutions were diluted to 150 mM for use in Cas13a activity assays. Guide design Guides were selected using a broad search process followed by down selection. Each genome was initially aligned and trimmed such that only fragment 11 (between nucleotide position 22,389 and 24,230 in SARS-CoV-2 genome Hu1 21 ) remained. Genomes were grouped into sets labeled by a color for easier visualization: black (Human SARS-CoV-2/ Wuhan-Hu1), dark red (BANAL-236, RaTG13, STT182, MP789, Guangxi-P4L), light blue (RsYN04), green (RacCS203, B247), dark blue (RsYN03), yellow (RpYN06, CoVZC45), and purple (Rc-o319). The guide (crRNA) nomenclature follows the color-based pattern with names of the guides composed of the abbreviated color name corresponding to their intended target (black = blk, dark red = dkrd, light blue = ltbl, green = grn, dark blue = dkbl, yellow = ylw and purple = prpl) and a consecutive number assigned by the software. For guide selection, the above sets were used for assigning inclusion and exclusion genomes. Separately, for the inclusion and exclusion sets, alignments were produced using MAFFT (v7.490) 22 . RPrimer functions 23 were used to read each alignment, produce a consensus profile, and possible guides were listed with a length of 28 base pairs using default parameters. Proposed guides unique to each set were identified such that oligos where only guides with ≤ 1 mismatch with set A (the inclusion set) and ≥ 4 mismatches for set B (the exclusion set) were selected. In each case, except for the near neighbors (relative to SARS-CoV-2/Wuhan-Hu1/black group) of the dark red genome set, guides were designed such that Wuhan-Hu1 was used as the exclusion set, and the inclusion genomes (e.g., RpYN06 and CoVZC45 for yellow) were to be detected; however, the extent of mismatch number and other features of the proposed guides with respect to other genomes not included in set A or B were not considered. Following selection of guides from the compiled list, grep (a standard Unix command-line utility) was used to double check mismatch number against the selected genome fragments. In addition to the mismatch number between the guide and the target sequences the interquartile range (IRQ) for a set of mismatch positions corresponding to a particular guide/target pair was calculated. As described in our previous work the IRQ value reflects the uniformity of distribution of mismatches across the length of the spacer. 20 The quartiles Q1 and Q3 were determined and IRQ was obtained by subtracting Q1 value from Q3 (IQR = Q3 – Q1). Values of IRQ close to 14 indicate a uniform distribution of mismatches along the spacer while values much lower than 14 correspond to mismatches arranged as a single cluster and values much higher than 14 correspond to mismatches arranged in two separate clusters. IRQ values are not obtainable for zero mismatches, one mismatch, and > 8 mismatches. The guide sequences for testing were down selected from the full list based on the number of mismatches and the IRQ values. The ten guides for specific SARS-CoV-2 detection were selected from among the sequences with no mismatches with Wuhan-Hu1 target and the highest number of mismatches with all other targets (Supplementary Table S1 ). Four guides for “dark red” and “purple” targets and five guides for the other target groups were selected as the zoonotic coronavirus targets. The guides were selected based on the highest number of mismatches and a value of IQR as far from 14 as possible for the pairings with the exclusion group targets (Supplementary Table S2 ). The guide sets used for the final modified SHERLOCK assay were selected based on the in-vitro testing results described further in the manuscript and took additional criteria into account (e.g., the number of mismatches needed for target exclusion was increased to > 6). crRNA synthesis The crRNA synthesis method described below has been described and validated in our two prior studies 24 , 25 . Sequences of the DNA oligonucleotides encoding crRNAs were designed by adding the variable spacer sequences to the 5’ end of the backbone sequence (direct repeat sequence) and T7 polymerase promotor sequence to the 3’ end of the backbone as reported previously (Supplementary Materials: Sequences) 24 . The crRNA molecules were obtained by conducting in vitro transcription of synthetic DNA oligonucleotides The oligonucleotides were purchased from IDT and listed in Supplementary Materials: Sequences. In vitro transcription was done using the HiScribe™ T7 Quick High Yield RNA Synthesis Kit. The individual transcription reactions were performed in 25 µL of total volume. This included 0.5 µL of 100 µM T7 forward primer, 1.5 µL of 100 µM crRNA-encoding DNA oligonucleotide, 1.25 µL of T7 RNA polymerase, 9.25 µL of 2x NTP buffer and 12.5 µL of nuclease-free ddH2O. The reactions were carried out for 2 h at 37°C. The obtained crRNAs were used in Cas13a activity assays without additional purification. Cas13a activity assays for testing crRNA performance The Cas13a activity assays detailed below have been developed and described in our previous publications and were used in this work with minor modifications. 20 , 25 To determine the efficacy of each crRNA, Cas13a nuclease activity assays were conducted using Cas13a enzyme from L. wadei 26 which was synthesized and purified by GenScript Biotech (Piscataway, NJ). Depending on the number of crRNA and RNA targets run at the same time the assay was conducted using an Echo 525 acoustic liquid handler (Beckman Coulter, Indianapolis, IN) using the Plate Reformat software provided by the manufacturer as described earlier 25 or performed manually. The Cas13a enzyme was stored and diluted using the storage buffer (50 mM Tris-HCl, 600 mM NaCl, 5% Glycerol, 2 mM DTT, pH 7.5). Each nuclease activity assay was performed in 20 µL reaction that included 1 µL of 1 µM Cas13a, 1 µL of 2 µM RNase alert v.2 (from RNaseAlert™ QC System v2, ThermoFisher), 17.2 µL of nuclease assay buffer (40 mM Tris-HCl, 60 mM NaCl, 6m M MgCl2, pH 7.3), 0.4 µL of crRNA (from unpurified transcription reaction) and 0.4 µL of 30 mM target RNA. For each crRNA a total of six reactions were set up, with three target negative reactions and three target positive wells. First, a master mix containing all reaction components except for the crRNA and target RNA were distributed to a 384 well assay plate. A total volume of 19.2 µL of the master-mix was transferred to each well. Next, 0.4 µL crRNAs were transferred to the wells containing the master-mix in such a way that each crRNA was added to 6 subsequent wells in the reaction plate. Finally, 0.4 µL of the target RNAs were added to three of the wells for each crRNA. The Cas13a reaction plates were spun briefly in a centrifuge at approximately 1500 x g to bring all the liquid to the bottom of the wells and remove air bubbles. Immediately after spinning, the reaction plates were sealed using MicroAmp sealers. The plates were incubated without shaking in SpectraMax M3 plate reader (Molecular Devices, San Jose, CA) at 37°C and fluorescence was read from the bottom of the wells every 5 minutes for 2 hours using excitation at 490 nm, emission at 520 nm with auto cutoff on and PMT gain set to “medium” and 6 flashes per read and carriage speed set to “normal”. The integrated background corrected final fluorescence values reflecting the Cas13a RNase activation for each of the crRNAs was calculated by subtracting the sum of averages of fluorescence measured for template negative samples over the course of the experiment (25 measurements) from sum of averages for template positive samples. Each of the tested crRNAs was classified as positive or negative based on the corrected integrated fluorescent signal, i.e. background subtracted, relative to the signal obtained for the perfectly matching target (target with zero mismatches). The assay was considered positive when the signal was equal or higher to 20% of the reference and negative for signal below 20%. Integrated modified SHERLOCK assay Detection of DNA and RNA version of the targets was conducted using two step assay broadly based on SHERLOCK detection method 27 . The assay was performed in two steps: step one, recombinase polymerase amplification (RPA) based target amplification combined with reverse transcription for RNA targets and step two: T7 based in vitro transcription combined with Cas13a activity assay. The RPA amplification step was performed in 10.8–11.3 µL total volume using TwistAmp® Liquid Basic kit (TwistDx Limited, Maidenhead, UK). The initial reaction mix was set up by combining 5 µL of the 2x TwistAmp reaction buffer, 1.8 µL of 10 mM dNTP mix (New England Biolabs), 1 µL TwistAmp 10x basic E mix, 0.5 µL of 10 µM of each amplification primer, 0.5 µL of Superscript III (ThermoFisher) and 0.5 µL of TwistAmp core reaction mix. In case of DNA templates, the Superscript III was omitted and replaced by nuclease-free water. The reagents were mixed by tube inversion or pipetting up and down. Next, 0.5 µL of magnesium acetate (MgOAc) and 1 µL of the template (target) DNA or RNA were added in separate drops to the lid of the tube. The tube was closed, and reagents mixed by inverting the tube 6 times. The reaction mixture was subsequently incubated at 37°C for one hour. The amplified targets were immediately used for detection in Cas13a assay (step two) or stored frozen at -80°C. The combined T7 based transcription and Cas13 detection was carried out in 30 µL reaction that included 1.25 µL of 1 µM Cas13a, 1.25 µL of 2 µM RNase alert v.2, 0.5 µL of unpurified crRNA rection mix, 0.25 µL of T7 RNA polymerase (from HiScribe™ T7 Quick High Yield RNA Synthesis Kit), 5 µL of 2x NTP buffer (from HiScribe™ T7 Quick High Yield RNA Synthesis Kit), 20.5 µL of nuclease assay buffer (40 mM Tris-HCl, 60 mM NaCl, 6m M MgCl 2 , pH 7.3) and 3 µL of the amplified target obtained in step one. For each detection six reaction were run including 3 negative controls (with TE buffer added in place of the amplified target) and 3 detection reactions. The reactions were incubated for 2 hours at 37°C in 384 reaction plates and read using the plate reader and standard Cas13a assay procedures as described above. Dataset, data processing, and feature extraction The data processing, and feature extraction methods described below have been previously published previously in our earlier work. 20 A dataset was constructed based on the results of the assays testing the performance of all the tested crRNAs with a panel of targets containing 13 coronavirus sequences representing SARS-CoV-2 and a several clades of nearest neighbor zoonotic coronaviruses. Each data entry included a list of positions of mismatches between the crRNA spacer and the corresponding target sequence together with a fluorescent signal obtained in the Cas13a activity assay using this crRNA spacer/target combination. To identify the mismatch positions the target sequences (converted to DNA sequence) and reverse complements of the crRNA spacers (also converted to DNA sequences) were compared. Mismatches were identified by applying binary labels for match/mismatch for each base and each spacer/target pairing. For each of the dataset entries 22 features were extracted or calculated from the mismatch data and target sequences as described previously 20 . The features most relevant for this work include: “n” – the total number of mismatches between the crRNA spacer (guide sequence) and the target sequences, “mean” – the mean value calculated for the spacer positions of all mismatches, “IQR” (interquartile range) – the difference between Q1 and Q3 quartile values of all mismatch positions for the spacer sequence, “min” – spacer position of a mismatch nearest to the crRNA hairpin (5’ end of crRNA) and “max” - spacer position of a mismatch nearest to the 3’ end of crRNA. The results of the Cas13a activity assays using a particular crRNA spacer/target combination were designated as either positive or negative based on the following criteria: a sample was evaluated as negative if the cumulative fluorescent, background subtracted, signal was less than 20% of the maximum signal obtained for the crRNA assay with target with no mismatches, and positive if it was greater than or equal to 20% of the maximum signal. Models and feature importance ranking Models were built to classify the spacer/target combinations as producing positive or negative assay outcomes as described earlier 20 . Rule-based models such as RuleFit use groupings (ensembles) of linear models to build either classification or regression predictions that are comparable in accuracy as the best alternatives 28 . However, their main advantage is in their interpretability, as each rule in the ensemble is a simple statement related to the individual features in the input dataset. This property of RuleFit allows for clear ranking of the relative importance of each feature and allows to better understand their data and the predictions. The classification model was generated in R using the Tidymodels series of packages 29 . Rule based Learning Ensembles (RuleFit) were assembled with the XRF package 30 . The number of trees contained in the ensemble was set to 2, maximum depth of the tree was set to 3, and the L1 regularization parameter was set to 0.01; all other parameters were set to defaults. Results Testing of the crRNA performance. The guides selected by the machine learning model were subsequently tested against the target analytes using a standard Cas13a activity assay with 1.5 nM final target concentration and unpurified in-vitro transcribed crRNAs in 20 µL total reaction volume as described in the methods section. We present the separate results focusing on each specific clade (i.e. color group), and the capability of each guide to maintain specificity within its group. A summary of the results is available in Fig. 3 with the detailed results available in the supporting information. Black group: SARS-CoV-2/Wuhan-Hu-1 specific The total of ten crRNAs specific for SARS-CoV-2 (black group) were selected for testing of their performance and specificity. Only guide sequences with zero mismatches to SARS-CoV-2 target and four or more mismatches to zoonotic targets were selected for this group. Each of the crRNAs was tested with SARS-CoV-2 and all 12 zoonotic targets (Detailed results available in Supplementary Figure S2 and Supplementary Table S1 ). All ten crRNAs showed good activity for SARS-CoV-2 (Wuhan-Hu-1) target, however majority of them also showed activity with several of the nearest neighbor targets belonging to “dark red” target group. All the crRNAs except crRNA blk-4 and blk-7 were positive for BANAL-236 target and all but crRNAs blk-4, blk-6 and blk-7 were positive for RShSTT182 (which are the two closest neighbors tested). In addition, crRNAs blk-8 and blk-9 were positive for RaTG13 and crRNA blk-9 was also additionally positive for MP789. The only two crRNAs which were positive exclusively for SARS-CoV-2 (Wuhan-Hu-1) target were blk-4 and blk-7 (Supplementary Table S1 ). Dark red group Four crRNAs specific for “dark red” group of targets (BANAL-236, RaTG13, RShSTT182, MP789, and Guangxi-P4L) were tested: dkrd-1, dkrd-2, dkrd-3, and dkrd-5 (Detailed results available in Supplementary Figure S3 and Supplementary Table S2 ). Due to a very high similarity of the targets from this group to SARS-CoV-2/Wuhan-Hu-1 the guide sequences for this crRNA group were designed using relaxed criteria without defining SARS-CoV-2/Wuhan-Hu-1 as the exclusion group (see Methods for details). As a result, all four of the tested crRNAs have between zero and two mismatches with SARS-CoV-2/Wuhan-Hu-1 target and were found positive for both “dark red” group of targets and SARS-CoV-2/Wuhan-Hu-1. Also, only the dkrd-1 and dkrd-2 crRNAs were found exclusively specific for “dark red” and SARS-CoV-2/Wuhan-Hu-1 targets. The crRNAs dkrd-3 and dkrd-5 were additionally found positive for targets belonging to the “yellow” group (SL-CoVZC45 and RpYN06). Light blue group Five crRNAs specific for the light blue target (RsYN04) were tested: ltbl-21, ltbl-122, ltbl-127, ltbl-175, ltbl-177 (Detailed results available in Supplementary Figure S4 and Supplementary Table S2 ). Only one crRNA (ltbl-177) was found to be completely specific and positive exclusively for the “light blue” target. The ltbl-21 was positive for the light blue target (RsYN04) and just above the threshold of positivity (22%) for RaTG13 (one of the “dark red” targets). The ltbl-122 crRNA failed to produce meaningful signal for any of the tested targets. One crRNA (ltbl-175) was positive for the “light blue” (RsYN04) and “dark blue” (RsYN03) targets and another one (ltbl-127) was found positive for the “light blue” (RsYN04) and two “yellow” (SL-CoVZC45 and RpYN06) targets. Green group Five crRNAs designed for detection of the “green” targets (BANAL-247 and RacCS203) were tested: grn-60, grn-87, grn-91, grn-97, and grn-133 (Detailed results available in Supplementary Figure S5 and Supplementary Table S2 ). Two crRNAs (grn-87 and grn-97) were found to be positive exclusively for green targets. The other two crRNAs (grn-60 and grn-91) were found positive only for one of the “green” targets (RacCS203). Finally, the crRNA grn-133 was found positive for both of the “green” targets and at the threshold of positivity (20%) for the “dark blue” (RsYN03) target. Dark blue group Five crRNAs designed for detection of the “dark blue” target (RsYN03) were tested: dkbl-24, dkbl-25, dkbl-64, dkbl-116, and dkbl-150 (Detailed results available in Supplementary Figure S6 and Supplementary Table S2 ). Only one of the tested crRNAs (dkbl-24) was found specific for the “dark blue” target. Two crRNAs (dkbl-64, dkbl-116) detected the “dark blue” target and one or both “yellow” targets (SL-CoVZC45 and RpYN06). The dkbl-25 crRNA detected one of the “dark red” targets (Guangxi-P4L) in addition to the “dark blue” target. The dkbl-25 was the least specific of the tested crRNA and generated signal above the positivity threshold for “dark blue” target, four “dark red” targets (RaTG13, RShSTT182, MP789, and Guangxi-P4L), both “yellow” targets and the “purple” target. Yellow group Five crRNAs designed for detection of the “yellow” targets (SL-CoVZC45 and RpYN06) were tested: ylw-45, ylw-75, ylw-83, ylw-86, and ylw-90 (Detailed results available in Supplementary Figure S7 and Supplementary Table S2 ). While all the tested crRNAs gave positive signal with both “yellow” targets, only one of them (ylw-83) was specific to these targets only. Three crRNAs (ylw-45, ylw-75, and ylw-90) were additionally positive for the “dark blue” target. Two of these crRNAs (ylw-45 and ylw-90) were also additionally positive for some of the “dark red” targets: Guangxi-P4L in case of ylw-45 and MP789 in case of ylw-90. The ylw-75 crRNA was found to be positive for the “green” target (RacCS203) in addition to “dark blue” and “yellow”. Finally, the ylw-86 crRNA was positive for both “yellow” targets and one of the “dark red” targets (MP789). Purple group Four crRNAs designed for detection of the “purple” target (Rco319) were tested: prpl-6, prpl-23, prpl-54, and prpl-105 (Detailed results available in Supplementary Figure S8 and Supplementary Table S2 ). All of the tested crRNAs produced strong signal for the “purple target”, hover only one crRNA (prpl-6) was specific to purple target. In addition to “purple” the prpl-23 crRNA was positive for “dark blue” target, the prpl-54 crRNA was positive for one of the “dark red” targets (RShSTT182) and the prpl-105 crRNA was positive for one of the “yellow” targets (SL-CoVZC45). Performance of the RuleFit based predictive model The RuleFit algorithm-based predictive model was previously developed using experimental data to predict crRNA performance based on the number and distribution of the mismatches between guide and target sequences 20 . The RuleFit classifier provides an initial estimate for whether a guide-target pair will result in a positive signal, i.e. >20% of max signal. That algorithm was used as is to select initial guide sequences for this study. To design guides which would generate signal with the inclusion targets yet not to detect the exclusion group the threshold was set at 4 mismatches or greater with the exclusion group targets. Unexpectedly, the in vitro testing results showed that multiple guide-target pairings with up to 6 mismatches could produce strong signal (e.g. 88% of max signal for SARS-CoV-2 specific guide black-9 with RshSTI182 target sequence). The model was retrained using the in vitro data obtained in this study and performance of the modified model was assessed again. The confusion matrix (Fig. 4 , panel A) compares the predictions of the updated model with the actual experimental data. The model shows an overall accuracy of 94%. Consistent with previous design rules, the feature with the overall highest global impact on the predictions was the total mismatch count (n). Increasing the threshold value to n ≥ 6 allow correct classification of 90% of the pairings. In contrast to the original findings where IQR was determined to be the second most important feature, after retraining the model, the “mean” feature replaced it as the second most important feature with IRQ moving to the third position (Fig. 4 , panel B). The significance of the “mean” feature suggests that mismatch positions closer to the 3’ end of the guide are less disruptive, which are in line with the earlier observations that shortening the guide sequence by up to seven nucleotides from the 3’ end has little impact on crRNA performance 31 . To determine the effect of shortening the guide sequence on the model predictions the model was retrained using a data set generated using just the first 20 positions of the guide/target pairing. As a result, it was found that for the shortened guides the threshold mismatch n value for 90% efficiency (Supplementary Figure S9, panel A) of predicting negative outcome dropped from 6 to 4 within the test set and the importance of the number mismatches for predictions increased significantly over other features (Supplementary Figure S9, panel B), with percentage of decisions explained by the number of mismatches increasing from 47–70%. Limit of detection (LOD) determination for SARS-CoV-2 DNA target for the integrated assay. A series of seven 10-fold dilutions of the Wuhan-Hu-1 synthetic DNA target ranging from 250 pM to 250 aM and seven 10-fold dilutions of synthetic RNA target ranging from 1.5 nM to 1.5 fM were prepared and tested with the two step SHERLOCK type assay detailed in the methods. The target preparations were amplified in the first step of the assay using primer pair specifically designed for amplification of Wuhan-Hu-1 synthetic target. The crRNAs used for the Cas13a part of the assay were an equimolar mixture of two Wuhan-Hu-1 specific crRNAs black-4 and black-7. The negative controls (no target control – NTC) was a Cas13a assay with TE buffer in place of the amplified target and positive control was a Cas13a assay with 1.5 nM Wuhan-Hu-1 RNA target. The results of the experiments are shown in Fig. 5 . The LOD for DNA targets was determined to be between 25 fM and 2.5 fM while the LOD using RNA targets was determined to be between 150 fM and 15 fM. The sensitivity of the assay for specific detection of Wuhan-Hu-1 synthetic target using a mixture of black-4 and black-7 crRNAs is approximately 6 times higher for DNA templates compared to RNA templates. The loss of sensitivity when using RNA templates seems to be the result of including of reverse transcription in the first step of the assay. Detection of mixed RNA targets and performance of the assay for discrimination of SARS-CoV-2 and its zoonotic near neighbors. With the intent of testing the inhibitory effect of non-target RNAs inclusion in the sample matrix we conducted a mixed RNA pool test. Results of the experiment testing detection of RNA targets in the mixed sample are shown in the Supplementary Figure S10. The experiment was conducted using a sample containing mixture of all 13 synthetic targets tested in this study at 1.5 nM final mixed target concentration. A total of eleven Cas13 assays were conducted using selected individual crRNAs designed for detection of SARS-CoV-2 and 6 distinct clades of zoonotic viruses. The following crRNAs were used in this assay black-7 (specific for Wuhan-Hu-1 target), dkrd-1, dkrd-2 (specific for both Wuhan-Hu-1 and “dark red” zoonotic clade), dkbl-24 (specific for “dark blue” zoonotic clade), grn-87, grn-133 (mostly specific for “green” zoonotic clade), ltbl-127 (specific for “light blue” and “yellow” zoonotic clades), ltbl-177 (specific for “light blue” zoonotic clade), prpl-6 (specific for “purple” zoonotic clade), ylw-75 (specific for “yellow” and “dark blue” zoonotic clades) and ylw-86 (specific for “yellow” zoonotic clade and cross reacting with one of the “dark red” targets). All of the detection assays were positive based on the 20% maximum signal threshold with the lowest signal (30%) obtained for ylw-86 and highest (100%) for prpl-6 crRNA. Duplex assay for discrimination of SARS-CoV-2 and zoonotic near neighbors As a final proof of concept, we aimed to demonstrate the capability to distinguish between a specific target (SARS-CoV-2) and its near neighbors using a limited set of detection channels. The intent being the capability to utilize a simple assay to help distinguish known specific target from natural near-neighbors or engineered variants. The two-step modified SHERLOCK assay was conducted using the following RNA targets: Wuhan-Hu-1 (SARS-CoV-2) and seven targets representing all zoonotic near neighbor clades: BANAL-236 and Guangxi-P4L (dark red), RsYN04 (light blue), RacCS203 (green), RsYN03 (dark blue), RpYN06 (yellow) and Rc-o319 (purple). The synthetic RNA targets at two different concentrations were amplified using RT/RPA and each of the amplified targets were used for two T7/Cas13a detection assays one using a SARS-CoV-2 specific crRNA mixture (blk-4 and blk-7) and the other using zoonotic-near-neighbor-detection mixture containing 10 distinct crRNAs and designed to detect both SARS-CoV-2 and closely related sarbecoviruses (dkrd-1, dkrd-2, ltbl-127, ltbl-177, grn-87, grn-133, dkbl-24, ylw-75, ylw-86, prpl-6). The results of this assay are summarized in the Fig. 6 and Supplementary Figure S11. The results show that the SARS-CoV-2 specific reaction was positive for Wuhan-Hu-1 but also positive for its nearest zoonotic neighbor BANAL-236 but with much lower fluorescent signal. The other zoonotic targets were negative in this assay. For the assay using zoonotic-near-neighbor-detection crRNA mixture all targets were found positive with varying signal levels. Taking both assay into account it was possible to distinguish between three cases: (1) SARS-CoV-2 positive (strongly positive SARS-CoV-2 assay and positive zoonotic-near-neighbor-detection assay with significantly lower fluorescent signal), (2) BANAL-236 positive (both assays positive with comparable signal levels) and (3) zoonotic-near-neighbor positive (negative SARS-CoV-2 assay and positive zoonotic-near-neighbor-detection assay). Conclusions This study demonstrates the feasibility of developing detection assays capable of differentiation of targets of varying degree of variability with small number of detection channels using CRISPR/Cas based RNA/DNA target detection coupled with guide sequence optimization with machine learning model. In this work we used a model system in which the aim was to distinguish RNA target sequences representing SARS-CoV-2 (specific target) from a group of closely related sarbecoviruses. The example was chosen due to its relevance, yet the developmental pipeline could be applied to other viruses or systems of interest. The RuleFit based machine learning model which was previously trained on Lassa virus (LASV) sequences was used for initial selection of guide sequences for detection of the receptor binding domain of spike gene. This region of genome is known for relatively high level of divergence between SARS-CoV-2 and zoonotic near neighbors compared to the adjacent sequences. In vitro verification of the actual ranges of specificity of these guide sequences have shown that the model needed further optimization. We found out that the four mismatch threshold between guide and target as determined using LASV dataset is not sufficient to avoid cross-reactivity with targets belonging to the exclusion groups. Retraining the model using the new in vitro data increasing the threshold to 6 mismatches allowed the model to achieve high level of accuracy. This final result highlights, the ubiquitous by now, knowledge that the quality of results from any machine learning model is dependent on the quality of the input data. While the LASV dataset was a good starting place to train the RuleFit model, use of the specific sarbecovirus training data further improved the results. Novel applications would therefore be able to use existing guide rules (e.g. LASV or sarbecovirus) for immediate results, while then retraining their model for subsequent optimization. Our initial testing using modified two-step SHERLOCK assay taking advantage of simultaneous reverse transcription and RPA amplification followed by T7 based transcription and Cas13a detection shows that the assay has the sufficient sensitivity to detect the virus in concentrations present in the nasal wash during the acute phase of the COVID infection 32 (LOD for RNA targets between 150 fM and 15 fM equivalent to target molecule concentrations between ~ 2.5x10^4 and ~ 2.5x10^3 copies/µL). The optimized guide sequence sets allowed us to build a two-channel detection assay which can not only detect SARS-CoV-2 but, at the same time, indicate the presence of one of the near-neighbor sarbecoviruses and distinguish situations in which both or just one of these (groups) of targets is present in the sample. This type of assays may be extremely useful for detection new zoonotic spillovers and use of genetically modified pathogens. While we would expect novel zoonotic, or even manmade, variants to be uncommon, our approach provides an early detection approach without massively expanding the assay requirements and costs. Upon a positive result for variants, the sample could be more thoroughly characterized for a more precise determination and the proper course of action. Declarations Competing Interests TAL and DAS are listed as inventors on a patent application for methods described herein. The remaining authors declare no competing interests. Funding declaration This work was funded by the Defense Threat Reduction Agency (HDTRA1240013) and Office of Naval Research Base funding to the U.S. Naval Research Laboratory. Author Contribution DAS and TAL conceived the study, DAS, TAL and CMG designed the experimental protocols, TAL performed the experimental part of the study, DAS, TAL, SAD, ZTG and CMG – analyzed the obtained experimental data, SND performed statistical and machine learning analysis of crRNA guide performance using RuleFit model, TAL wrote the first version of the manuscript. DAS, SAD, ZTG and CMG reviewed and made edits to the manuscript, TAL, ZTG and CMG prepared the manuscript figures, DAS obtained the funding for this work. All authors read and approved the submitted version of the manuscript. Acknowledgement This manuscript was approved for public release by DTRA and U.S. Naval Research Laboratory with unlimited distribution. Data Availability All data and code used in this study are available through the GitHub repository at https://github.com/NRL-CRISPR/CRISPR-rules. References Li, H. Y. et al. A qualitative study of zoonotic risk factors among rural communities in southern China. Int. Health . 12 , 77–85. 10.1093/inthealth/ihaa001 (2020). Chomel, B. B., Belotto, A. & Meslin, F. X. Wildlife, exotic pets, and emerging zoonoses. Emerg. Infect. Dis. 13 , 6–11. 10.3201/eid1301.060480 (2007). Field, H. E. Bats and emerging zoonoses: henipaviruses and SARS. Zoonoses Public. Health . 56 , 278–284. 10.1111/j.1863-2378.2008.01218.x (2009). Milne-Price, S., Miazgowicz, K. L. & Munster, V. J. The emergence of the Middle East respiratory syndrome coronavirus. Pathog Dis. 71 , 121–136. 10.1111/2049-632X.12166 (2014). Crits-Christoph, A. et al. Genetic tracing of market wildlife and viruses at the epicenter of the COVID-19 pandemic. Cell 187, 5468–5482 e5411, (2024). 10.1016/j.cell.2024.08.010 Mozer, A. & Prost, S. An introduction to illegal wildlife trade and its effects on biodiversity and society. Forensic Sci. International: Anim. Environ. 3 , 100064. https://doi.org/10.1016/j.fsiae.2023.100064 (2023). Hughes, A. et al. Determining the sustainability of legal wildlife trade. J. Environ. Manage. 341 , 117987. 10.1016/j.jenvman.2023.117987 (2023). Ahteensuu, M. S. & Biology Genome Editing, and the Risk of Bioterrorism. Sci. Eng. Ethics . 23 , 1541–1561. 10.1007/s11948-016-9868-9 (2017). Sabra, D. M., Krin, A., Romeral, A. B., Friess, J. L. & Jeremias, G. Anthrax revisited: how assessing the unpredictable can improve biosecurity. Front. Bioeng. Biotechnol. 11 , 1215773. 10.3389/fbioe.2023.1215773 (2023). Gerstein, D. M., Espinosa, B. & Leidy, E. N. Emerging technology and risk analysis - synthetic pandemics (Homeland Security Operational Analysis Center operated by the RAND Corporation, 2024). Kamalrathne, T., Amaratunga, D., Haigh, R., Kodituwakku, L. & Rupasinghe, C. Epidemic and Pandemic Preparedness and Response in a Multi-Hazard Context: COVID-19 Pandemic as a Point of Reference. Int. J. Environ. Res. Public. Health . 21 10.3390/ijerph21091238 (2024). Liu, Q. et al. Advances in the application of molecular diagnostic techniques for the detection of infectious disease pathogens (Review). Mol. Med. Rep. 27 10.3892/mmr.2023.12991 (2023). Patchsung, M. et al. Clinical validation of a Cas13-based assay for the detection of SARS-CoV-2 RNA. Nat. Biomed. Eng. 4 , 1140–1149. 10.1038/s41551-020-00603-x (2020). Green, C. M. et al. Quantum Dot-Based Molecular Beacons for Quantitative Detection of Nucleic Acids with CRISPR/Cas(N) Nucleases. ACS Nano . 16 , 20693–20704. 10.1021/acsnano.2c07749 (2022). Gootenberg, J. S. et al. Multiplexed and portable nucleic acid detection platform with Cas13, Cas12a, and Csm6. Science 360 , 439–444. 10.1126/science.aaq0179 (2018). Guo, X. et al. Transcriptome-wide Cas13 guide RNA design for model organisms and viral RNA pathogens. Cell. Genomics . 1 , 100001. https://doi.org/10.1016/j.xgen.2021.100001 (2021). Wessels, H. H. et al. Massively parallel Cas13 screens reveal principles for guide RNA design. Nat. Biotechnol. 38 , 722–727. 10.1038/s41587-020-0456-9 (2020). Molina Vargas, A. M. et al. New design strategies for ultra-specific CRISPR-Cas13a-based RNA detection with single-nucleotide mismatch sensitivity. Nucleic Acids Res. 52 , 921–939. 10.1093/nar/gkad1132 (2024). Durán-Vinet, B. et al. CRISPR-Based Environmental Biosurveillance Assisted via Artificial Intelligence Design of Guide-RNAs. Environ. DNA . 7 , e70105. https://doi.org/10.1002/edn3.70105 (2025). Leski, T. A. et al. Machine learning for design of degenerate Cas13a crRNAs using lassa virus as a model of highly variable RNA target. Sci. Rep. 13 , 6506. 10.1038/s41598-023-33494-4 (2023). Temmam, S. et al. Bat coronaviruses related to SARS-CoV-2 and infectious for human cells. Nature 604 , 330–336. 10.1038/s41586-022-04532-4 (2022). Katoh, K. & Standley, D. M. MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Mol. Biol. Evol. 30 , 772–780. 10.1093/molbev/mst010 (2013). Persson, S., Larsson, C., Simonsson, M. & Ellstrom, P. rprimer: an R/bioconductor package for design of degenerate oligos for sequence variable viruses. BMC Bioinform. 23 , 239. 10.1186/s12859-022-04781-0 (2022). Schultzhaus, Z., Wang, Z. & Stenger, D. Systematic analysis, identification, and use of CRISPR/Cas13a-associated crRNAs for sensitive and specific detection of the lcrV gene of Yersinia pestis . Diagn. Microbiol. Infect. Dis. 99 , 115275. 10.1016/j.diagmicrobio.2020.115275 (2021). Spangler, J. R., Leski, T. A., Schultzhaus, Z., Wang, Z. & Stenger, D. A. Large scale screening of CRISPR guide RNAs using an optimized high throughput robotics system. Sci. Rep. 12 , 13953. 10.1038/s41598-022-17474-8 (2022). Gootenberg, J. S. et al. Nucleic acid detection with CRISPR-Cas13a/C2c2. Science 356 , 438–442. 10.1126/science.aam9321 (2017). Kellner, M. J., Koob, J. G., Gootenberg, J. S., Abudayyeh, O. O. & Zhang, F. SHERLOCK: nucleic acid detection with CRISPR nucleases. Nat. Protoc. 14 , 2986–3012. 10.1038/s41596-019-0210-2 (2019). Friedman, J. H. & Popescu, B. A. Predictive learning via rule ensembles. Annals Appl. Stat. 2 , 916–954 (2008). Kuhn, M. & Wickham, H. Tidymodels: a collection of packages for modeling and machine learning using tidyverse principles. , (2020). https://www.tidymodels.org Friedman, J. H. & Popescu, B. E. Predictive learning via rule ensembles. annals Appl. Stat. 2 , 916–954 (2008). Abudayyeh, O. O. et al. RNA targeting with CRISPR-Cas13. Nature 550 , 280–284. 10.1038/nature24049 (2017). Savela, E. S. et al. Quantitative SARS-CoV-2 Viral-Load Curves in Paired Saliva Samples and Nasal Swabs Inform Appropriate Respiratory Sampling Site and Analytical Test Sensitivity Required for Earliest Viral Detection. J. Clin. Microbiol. 60 , e0178521. 10.1128/JCM.01785-21 (2022). Additional Declarations Competing interest reported. TAL and DAS are listed as inventors on a patent application for methods described herein. The remaining authors declare no competing interests. Supplementary Files CRISPRfordetectionofrangesofRNAtargetsSupplementaryTablesandSequences.xlsx CRISPRfordetectionofrangesofRNAtargetssupplementaryfigures.docx Cite Share Download PDF Status: Under Revision Version 1 posted Editorial decision: Revision requested 24 Nov, 2025 Reviews received at journal 20 Nov, 2025 Reviews received at journal 08 Nov, 2025 Reviewers agreed at journal 28 Oct, 2025 Reviewers agreed at journal 22 Oct, 2025 Reviews received at journal 15 Aug, 2025 Reviewers agreed at journal 05 Aug, 2025 Reviewers invited by journal 02 Aug, 2025 Editor assigned by journal 23 Jul, 2025 Editor invited by journal 11 Jul, 2025 Submission checks completed at journal 09 Jul, 2025 First submitted to journal 09 Jul, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7041916","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":496126656,"identity":"7b148cf4-1584-4188-8bbb-5d1b2026639c","order_by":0,"name":"Tomasz A. Leski","email":"data:image/png;base64,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","orcid":"","institution":"Naval Research Laboratory","correspondingAuthor":true,"prefix":"","firstName":"Tomasz","middleName":"A.","lastName":"Leski","suffix":""},{"id":496126657,"identity":"b8c61a51-a19c-4298-8ccc-d534dc3fe10c","order_by":1,"name":"Scott N. Dean","email":"","orcid":"","institution":"Naval Research Laboratory","correspondingAuthor":false,"prefix":"","firstName":"Scott","middleName":"N.","lastName":"Dean","suffix":""},{"id":496126658,"identity":"4aba2bea-ddb8-491a-8b5e-9de749d50952","order_by":2,"name":"Zachary T. Johnson","email":"","orcid":"","institution":"Naval Research Laboratory","correspondingAuthor":false,"prefix":"","firstName":"Zachary","middleName":"T.","lastName":"Johnson","suffix":""},{"id":496126659,"identity":"d8631875-0531-44fa-b006-e7798bd59dd6","order_by":3,"name":"Christopher M. Green","email":"","orcid":"","institution":"Naval Research Laboratory","correspondingAuthor":false,"prefix":"","firstName":"Christopher","middleName":"M.","lastName":"Green","suffix":""},{"id":496126660,"identity":"23128e8e-f927-4a93-967d-a5e6bfb2d4b4","order_by":4,"name":"Sebastián A. Díaz","email":"","orcid":"","institution":"Naval Research Laboratory","correspondingAuthor":false,"prefix":"","firstName":"Sebastián","middleName":"A.","lastName":"Díaz","suffix":""},{"id":496126661,"identity":"36758da9-a6dd-49fb-bb36-3e8985b9f4c1","order_by":5,"name":"David A. Stenger","email":"","orcid":"","institution":"Naval Research Laboratory","correspondingAuthor":false,"prefix":"","firstName":"David","middleName":"A.","lastName":"Stenger","suffix":""}],"badges":[],"createdAt":"2025-07-04 01:08:13","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7041916/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7041916/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":88662242,"identity":"4ef49c5d-1ef8-4cd9-815d-a9236c003bd4","added_by":"auto","created_at":"2025-08-08 21:32:53","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":25985,"visible":true,"origin":"","legend":"\u003cp\u003eCRISPR based detection system for discrimination between different ranges of closely related RNA targets.\u003c/p\u003e\n\u003cp\u003ePanel A: illustration of general idea of the two-channel detection system. Two separate CRISPR based detection reactions using different gRNA (crRNA) sequence sets are used for detection of discrimination of different target ranges present in a single sample. Panel B: the design of gRNA sets is based on the phylogenetic analysis of the target ranges. Panel C: an example output of the two-channel assay. In this example a narrow range of targets (SARS-CoV-2) and broad range of closely related targets (zoonotic sarbecoviruses) can be detected.\u003c/p\u003e","description":"","filename":"Picture1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7041916/v1/f98c6fb058878fc7cd4cdf4e.jpg"},{"id":88662696,"identity":"7da0140b-208b-4617-b4fc-b2f10c0c54c3","added_by":"auto","created_at":"2025-08-08 21:40:53","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":57304,"visible":true,"origin":"","legend":"\u003cp\u003eTarget and gRNA design and locations.\u003c/p\u003e\n\u003cp\u003ePanel A: the F11 region selected as a target region in the context of the SARS-CoV-2 genome. Panel B: the phylogenetic tree constructed using F11 region sequences from SARS-CoV-2 and closely related zoonotic sarbecoviruses. The arrows indicate the strains used for in vitro testing as representatives of the corresponding clades. The colors of the arrows correspond to the subgroups of targets as defined in this study. Panel C: locations of the gRNAs for each subgroup selected using machine learning based algorithm. These gRNAs were used for \u003cem\u003ein vitro\u003c/em\u003e testing.\u003c/p\u003e\n\u003cp\u003eThe panel B of this figure utilizes a phylogenetic tree reproduced with publishers permission from part of supplementary figure S2 originally published in Temmam, S. et al. Nature 604, 330-336, doi:10.1038/s41586-022-04532-4 (2022). The tree was minimally modified to indicate the sequences used as targets in this study and to define the color-coded groupings of related targets.\u003c/p\u003e","description":"","filename":"Picture2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7041916/v1/4f9c05886f8f80da5872ad29.jpg"},{"id":88662695,"identity":"79e7129b-cd19-46eb-833b-c5d5cbb2a491","added_by":"auto","created_at":"2025-08-08 21:40:53","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":26888,"visible":true,"origin":"","legend":"\u003cp\u003egRNA in vitro testing results.\u003c/p\u003e\n\u003cp\u003eAll the gRNA sequences were tested against SARS-CoV-2 and twelve zoonotic sarbecovirus RNA targets using Cas13a activity assay. The colored circles correspond to target groups and gRNAs optimized for specific detection of these target groups.\u003c/p\u003e","description":"","filename":"Picture3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7041916/v1/e21407c0cc8282e623f92f43.jpg"},{"id":88662244,"identity":"699abe98-9cca-4ed5-86c8-cc8f5aeab2cb","added_by":"auto","created_at":"2025-08-08 21:32:53","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":16098,"visible":true,"origin":"","legend":"\u003cp\u003eModel performance and the most important features\u003c/p\u003e\n\u003cp\u003eRuleFit classifier model performance. A – Confusion matrix showing percentages of actual assay outcomes versus outcomes produced by the classifier model, B – relative effect of features on model predictions.\u003c/p\u003e","description":"","filename":"Picture4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7041916/v1/ededac48834e207560eecdfa.jpg"},{"id":88662249,"identity":"265982de-d83d-49a5-83d4-59b5843b9196","added_by":"auto","created_at":"2025-08-08 21:32:53","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":31019,"visible":true,"origin":"","legend":"\u003cp\u003eDetermination of limit of detection for DNA and RNA targets.\u003c/p\u003e\n\u003cp\u003ePanel A: results of testing a series of 10-fold dilutions of Wuhan-Hu-1 synthetic DNA F11 target using modified SHERLOCK assay using a gRNA mix specific for SARS-CoV-2. Panel B: results of testing a series of 10-fold dilutions of Wuhan-Hu-1 synthetic RNA F11 target using modified SHERLOCK assay.\u003c/p\u003e","description":"","filename":"Picture5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7041916/v1/559fec707fa0735b01a716c9.jpg"},{"id":88662266,"identity":"aad9a998-09f7-4be7-8760-f6e3ed6afd35","added_by":"auto","created_at":"2025-08-08 21:32:53","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":20025,"visible":true,"origin":"","legend":"\u003cp\u003eTesting of two-channel modified SHERLOCK assay for detection and discrimination of SARS-CoV-2 and zoonotic near-neighbor sarbecoviruses.\u003c/p\u003e\n\u003cp\u003eThe heat maps show the intensity of the fluorescent signal for Wuhan-Hu-1 (SARS-CoV-2) and 12 zoonotic near neighbors for two different concentrations of synthetic F11 RNA targets: 1.5 nM (panel A) and 15 pM (panel B). The results allow for classifying the samples into three groups: 1) containing SARS-CoV-2, 2) containing BANAL-236 (the closest known SARS-CoV-2 neighbor in the terms of F11 region) and 3) all the remaining zoonotic near-neighbors.\u003c/p\u003e","description":"","filename":"Picture6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7041916/v1/e702814e1975157a5439a05c.jpg"},{"id":89063006,"identity":"b0bb82ac-3cc8-46bb-b563-b98cc7cabd4d","added_by":"auto","created_at":"2025-08-14 09:56:45","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1150474,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7041916/v1/f9f56760-1800-455f-82aa-8b08d96538af.pdf"},{"id":88662248,"identity":"1083d953-81d5-45a0-ad7d-d7ca4c5bf7bd","added_by":"auto","created_at":"2025-08-08 21:32:53","extension":"xlsx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":43311,"visible":true,"origin":"","legend":"","description":"","filename":"CRISPRfordetectionofrangesofRNAtargetsSupplementaryTablesandSequences.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7041916/v1/372596294c2b61e0fc8ef7e3.xlsx"},{"id":88663013,"identity":"71cc304e-e011-40d0-99ec-9197a2ef83ba","added_by":"auto","created_at":"2025-08-08 21:48:53","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":879351,"visible":true,"origin":"","legend":"","description":"","filename":"CRISPRfordetectionofrangesofRNAtargetssupplementaryfigures.docx","url":"https://assets-eu.researchsquare.com/files/rs-7041916/v1/0a12ecc3961903b6502b4e16.docx"}],"financialInterests":"Competing interest reported. TAL and DAS are listed as inventors on a patent application for methods described herein. The remaining authors declare no competing interests.","formattedTitle":"Discrimination of ranges of closely related RNA targets using CRISPR based detection assay developed using machine learning based optimization","fulltext":[{"header":"Introduction","content":"\u003cp\u003eThe emergence of infectious pathogens capable of causing global pandemics has accelerated in recent years, driven by several interconnected factors. The expanding human population, particularly in tropical regions, has led to encroachment into natural habitats, increasing interactions between humans and wildlife. This heightened contact elevates the risk of zoonotic diseases\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e,\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e. Several recently discovered viruses such as Hendra, Nipah, SARS-CoV, MERS-CoV and SARS-CoV-2 have been linked to wildlife reservoirs\u003csup\u003e\u003cspan additionalcitationids=\"CR4\" citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e. The thriving, largely illegal, and often unregulated wildlife trade further exacerbates the risk of zoonotic transmission and introduces new pathogens into human populations\u003csup\u003e\u003cspan additionalcitationids=\"CR6\" citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eBeyond natural zoonoses, there is a growing threat from engineered or synthetic pathogens used as biological weapons. Advancements in genetic engineering and synthetic biology have made development of new pathogens more accessible and affordable, raising concerns about the potential misuse of these technologies by state and non-state actors\u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e,\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e. The increasing availability of these technologies heightens the risk of engineered pathogens leading to potential \"synthetic pandemics\" in the coming decade\u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eEarly detection and identification of pathogens are crucial for effective strategies to mitigate both natural and man-made pandemic threats\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e. Recent advancements in molecular biology have significantly enhanced the sensitivity and specificity of pathogen detection methods, enabling rapid responses to infectious disease outbreaks\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e. Techniques such as polymerase chain reaction (PCR), various isothermal amplification methods, high-throughput sequencing, CRISPR-based detection, and immuno-detection technologies have been pivotal in this progress. However, a significant limitation of many current detection technologies is their narrow focus on one or a limited number of specific targets. This approach can lead to delays in identifying outbreaks caused by pathogens that are closely related but not identical to the intended targets, including engineered pathogens. CRISPR/Cas based assays, specifically the ones using Cas12 and Cas13 are capable of amplifying DNA or RNA target detection through the activation of their non-specific nuclease capability upon target binding. The SHERLOCK (Specific High-sensitivity Enzymatic Reporter unLOCKing) assay is the most well-known implementation of the Cas13 based assay and was clinically validated for detection of many different pathogens including SARS-CoV-2\u003csup\u003e13\u003c/sup\u003e. The assay readout usually relies on detection of fluorescent signal generated by digestion of RNA based molecular beacons by the activated non-specific nuclease. Both commercial and novel readout methodologies are being developed for improved detection or point of care applications\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e,\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e. As a result of these developments the design of proper guides has become a key focus of the field. While the majority of researchers have focused on guide design for cellular engineering (i.e. knockdowns)\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e,\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e, the work on methods of design of guides for RNA detection and its use in bio-surveillance has been also recently expanding \u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e,\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eThe goal of this study was to develop a CRISPR based assay capable of detecting a specific target and simultaneously distinguishing it from a broad range of closely related near-neighbor targets. The schematic representation of our assay with two detection channels is presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. To achieve this goal we needed to construct a developmental pipeline capable of efficiently designing appropriate RNA guide sets with the right degree of specificity that would allow for the distinction between a specific target (inclusion target set) and near neighbors (exclusion target set). In this work we used the SARS-CoV-2 virus and a collection of zoonotic sarbecoviruses closely related to SARS-CoV-2 as a model system. We applied a previously developed machine-learning-based model to design two types of guide sets for use in Cas13a-based (modified SHERLOCK) detection assays\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e. Our results were then used to retrain the model and resulted in better guide predictions. One guide set was able to specifically detect SARSCoV-2 without significantly cross-reacting with the near-neighbor species and second designed to detect several different clades of SARS-CoV-2 related zoonotic sarbecoviruses which represent the closest neighbors of SARS-CoV-2. The guide sequences were tested \u003cem\u003ein vitro\u003c/em\u003e, and a selected group of the guides was used to develop the final two-channel assay.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e All experimental procedures described in this manuscript were carried out in accordance with relevant guidelines and regulations including biosafety and chemical safety regulations. All nucleic acid sequences used were obtained from publicly available NCBI/GenBank and GISAID (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.gisaid.org/\u003c/span\u003e\u003cspan address=\"https://www.gisaid.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) database collections and no human subject research was conducted during this project. No materials classified as select agents were obtained or used in any experiments described in this manuscript.\u003c/p\u003e\u003cp\u003e\u003cb\u003eDesign of SARS-CoV-2 and other coronavirus target sequences and primers\u003c/b\u003e\u003c/p\u003e\u003cp\u003eCoronavirus genomes are known for their mosaic structure due to their high recombination rates. Temmam et al.\u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e identified several recombination breakpoints in SARS-CoV-2 genome which separate fragments of genome with different origin. This study focused on the fragment 11 as defined by Temmam, which codes for the C-terminal part of the Spike protein and more precisely the receptor binding domain (RBD). This part of the gene is one the most divergent among closest zoonotic near neighbors of SARS-CoV-2. In addition, SARS-CoV-2 contains a Furin cleavage site (FCS) in the middle of this region, a sequence which is absent from all known close neighbors of SARS-CoV-2. The target sequences chosen for this study included sequence of the ancestral SARS-CoV-2 strain Wuhan-Hu-1 and sequences from twelve near neighbor zoonotic isolates representing all major clades of SARS-CoV-2-related lineage of Sarbecoviruses. The phylogenetic relationship of the fragment 11 sequences for isolates selected for the study is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e and the information on these isolates: names, origin, accession numbers, lengths and starting as well as ending positions of the target sequences in the genomes of the coronavirus isolates are listed in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eOrigin and group assignment of target sequences\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"8\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eIsolate\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eOrigin\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eHost\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eAccession#\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eStart\u003c/p\u003e\u003cp\u003eposition\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eEnd\u003c/p\u003e\u003cp\u003eposition\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u003cp\u003eLength\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c8\"\u003e\u003cp\u003eGroup\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWuhan-Hu-1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eChina\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003ehuman\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eNC045512.2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e23310\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e23950\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e641\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eBlack\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBANAL-236\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eLaos\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cem\u003eRhinolophus marshalli\u003c/em\u003e (bat)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eEPI_ISL_4302647\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e23272\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e23900\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e629\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eDark red\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRaTG13\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eChina\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cem\u003eRhinolophus affinis\u003c/em\u003e (bat)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eEPI_ISL_402131\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e23307\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e23935\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e629\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eDark red\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRShSTT182\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCambodia\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cem\u003eRhinolophus shameli\u003c/em\u003e (bat)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eEPI_ISL_852604\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e23228\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e23856\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e629\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eDark red\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMP789\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eChina\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cem\u003eManis javanica\u003c/em\u003e (pangolin)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eMT121216.1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e23156\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e23784\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e629\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eDark red\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eGuangxi-P4L\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eChina\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cem\u003eManis javanica\u003c/em\u003e (pangolin)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eEPI_ISL_410538\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e23281\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e23909\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e629\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eDark red\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRsYN04\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eChina\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cem\u003eRhinolophus stheno\u003c/em\u003e (bat)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eEPI_ISL_1699444\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e23220\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e23842\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e623\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eLight blue\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBANAL-247\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eLaos\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cem\u003eRhinolophus malayanus\u003c/em\u003e (bat)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eEPI_ISL_4302648\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e23145\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e23767\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e623\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eGreen\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRacCS203\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eThailand\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cem\u003eRhinolophus acuminatus\u003c/em\u003e (bat)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eMW251308.1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e23189\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e23811\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e623\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eGreen\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRsYN03\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eChina\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cem\u003eRhinolophus sinicus\u003c/em\u003e (bat)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eEPI_ISL_1699443\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e23156\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e23784\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e629\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eDark blue\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSL-CoVZC45\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eChina\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cem\u003eRhinolophus pusillus\u003c/em\u003e (bat)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eMG772933.1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e23227\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e23855\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e629\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eYellow\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRpYN06\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eChina\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cem\u003eRhinolophus pusillus\u003c/em\u003e (bat)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eEPI_ISL_1699446\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e23237\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e23865\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e629\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eYellow\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRco319\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eJapan\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cem\u003eRhinolophus comutus\u003c/em\u003e (bat)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eLC556375.1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e23150\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e23781\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e632\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003ePurple\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"8\"\u003eThe panel B of this figure utilizes a phylogenetic tree reproduced with publishers permission from part of supplementary figure \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e originally published in Temmam, S. et al. Nature 604, 330\u0026ndash;336, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41586-022-04532-4\u003c/span\u003e\u003cspan address=\"10.1038/s41586-022-04532-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022). The tree was minimally modified to indicate the sequences used as targets in this study and to define the color-coded groupings of related targets.\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003e\u003cb\u003eRNA target synthesis\u003c/b\u003e\u003c/p\u003e\u003cp\u003eSynthetic double stranded DNA fragments used for synthesis of the RNA targets used in this study were ordered from Integrated DNA Technologies Inc. (IDT, Coralville, IA). The PCR primers pairs complementary to the ends of the synthetic fragments were designed for each target. T7 RNA polymerase promoter sequences were added to the 5\u0026rsquo; end of the forward primer in each primer pair. The PCR primers were also purchased from IDT. The sequences the synthetic DNA fragments and PCR primers are listed in Supplementary Material: Sequences.\u003c/p\u003e\u003cp\u003eThe target RNA molecules were produced using HiScribe\u0026trade; T7 Quick High Yield RNA Synthesis Kit (New England Biolabs, Ipswich, MA) as described in our previous work.\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e The synthetic DNA fragments of all coronavirus sequences described above were amplified, using the PCR primers, by FastStart Taq DNA polymerase kit (Millipore-Sigma, Burlington, MA) according to the manufacturer\u0026rsquo;s instructions. The forward PCR primers included T7 promotor sequences, which were incorporated into the amplicons, Supplementary Fig.\u0026nbsp;1. The transcription reactions were set up using 2 \u0026micro;L of unpurified DNA amplicon preparation, 2 \u0026micro;L of T7 RNA polymerase, 10 \u0026micro;L of 2x NTP buffer and 16 \u0026micro;L of nuclease-free ddH2O (30 \u0026micro;L of total reaction volume). The transcription reactions were incubated at 37\u0026deg;C for 2 h after which 5 \u0026micro;L Turbo DNAse (ThermoFisher, Grand Island, NY) and 15 \u0026micro;L of nuclease-free ddH\u003csub\u003e2\u003c/sub\u003eO were added (increasing the total volume to 40 \u0026micro;L) and incubated further 30 minutes at 37\u0026deg;C to remove the template DNA. The obtained transcript preparations were cleaned up using RNA Clean and Concentrator 25 kit (Zymo Research, Irvine, CA USA) according to the manufacturer instructions. The RNA concentration was determined using Qubit fluorometer and RNA BR (broad range) assay kit (ThermoFisher). The template solutions were diluted to 150 mM for use in Cas13a activity assays.\u003c/p\u003e\u003cp\u003e\u003cb\u003eGuide design\u003c/b\u003e\u003c/p\u003e\u003cp\u003eGuides were selected using a broad search process followed by down selection. Each genome was initially aligned and trimmed such that only fragment 11 (between nucleotide position 22,389 and 24,230 in SARS-CoV-2 genome Hu1 \u003csup\u003e21\u003c/sup\u003e) remained. Genomes were grouped into sets labeled by a color for easier visualization: black (Human SARS-CoV-2/ Wuhan-Hu1), dark red (BANAL-236, RaTG13, STT182, MP789, Guangxi-P4L), light blue (RsYN04), green (RacCS203, B247), dark blue (RsYN03), yellow (RpYN06, CoVZC45), and purple (Rc-o319). The guide (crRNA) nomenclature follows the color-based pattern with names of the guides composed of the abbreviated color name corresponding to their intended target (black\u0026thinsp;=\u0026thinsp;blk, dark red\u0026thinsp;=\u0026thinsp;dkrd, light blue\u0026thinsp;=\u0026thinsp;ltbl, green\u0026thinsp;=\u0026thinsp;grn, dark blue\u0026thinsp;=\u0026thinsp;dkbl, yellow\u0026thinsp;=\u0026thinsp;ylw and purple\u0026thinsp;=\u0026thinsp;prpl) and a consecutive number assigned by the software.\u003c/p\u003e\u003cp\u003eFor guide selection, the above sets were used for assigning inclusion and exclusion genomes. Separately, for the inclusion and exclusion sets, alignments were produced using MAFFT (v7.490)\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e. RPrimer functions\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e were used to read each alignment, produce a consensus profile, and possible guides were listed with a length of 28 base pairs using default parameters. Proposed guides unique to each set were identified such that oligos where only guides with \u0026le;\u0026thinsp;1 mismatch with set A (the inclusion set) and \u0026ge;\u0026thinsp;4 mismatches for set B (the exclusion set) were selected. In each case, except for the near neighbors (relative to SARS-CoV-2/Wuhan-Hu1/black group) of the dark red genome set, guides were designed such that Wuhan-Hu1 was used as the exclusion set, and the inclusion genomes (e.g., RpYN06 and CoVZC45 for yellow) were to be detected; however, the extent of mismatch number and other features of the proposed guides with respect to other genomes not included in set A or B were not considered. Following selection of guides from the compiled list, grep (a standard Unix command-line utility) was used to double check mismatch number against the selected genome fragments.\u003c/p\u003e\u003cp\u003eIn addition to the mismatch number between the guide and the target sequences the interquartile range (IRQ) for a set of mismatch positions corresponding to a particular guide/target pair was calculated. As described in our previous work the IRQ value reflects the uniformity of distribution of mismatches across the length of the spacer.\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e The quartiles Q1 and Q3 were determined and IRQ was obtained by subtracting Q1 value from Q3 (IQR\u0026thinsp;=\u0026thinsp;Q3 \u0026ndash; Q1). Values of IRQ close to 14 indicate a uniform distribution of mismatches along the spacer while values much lower than 14 correspond to mismatches arranged as a single cluster and values much higher than 14 correspond to mismatches arranged in two separate clusters. IRQ values are not obtainable for zero mismatches, one mismatch, and \u0026gt;\u0026thinsp;8 mismatches.\u003c/p\u003e\u003cp\u003eThe guide sequences for testing were down selected from the full list based on the number of mismatches and the IRQ values. The ten guides for specific SARS-CoV-2 detection were selected from among the sequences with no mismatches with Wuhan-Hu1 target and the highest number of mismatches with all other targets (Supplementary Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). Four guides for \u0026ldquo;dark red\u0026rdquo; and \u0026ldquo;purple\u0026rdquo; targets and five guides for the other target groups were selected as the zoonotic coronavirus targets. The guides were selected based on the highest number of mismatches and a value of IQR as far from 14 as possible for the pairings with the exclusion group targets (Supplementary Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe guide sets used for the final modified SHERLOCK assay were selected based on the \u003cem\u003ein-vitro\u003c/em\u003e testing results described further in the manuscript and took additional criteria into account (e.g., the number of mismatches needed for target exclusion was increased to \u0026gt;\u0026thinsp;6).\u003c/p\u003e\u003cp\u003e\u003cb\u003ecrRNA synthesis\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThe crRNA synthesis method described below has been described and validated in our two prior studies \u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e,\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e. Sequences of the DNA oligonucleotides encoding crRNAs were designed by adding the variable spacer sequences to the 5\u0026rsquo; end of the backbone sequence (direct repeat sequence) and T7 polymerase promotor sequence to the 3\u0026rsquo; end of the backbone as reported previously (Supplementary Materials: Sequences) \u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e. The crRNA molecules were obtained by conducting \u003cem\u003ein vitro\u003c/em\u003e transcription of synthetic DNA oligonucleotides The oligonucleotides were purchased from IDT and listed in Supplementary Materials: Sequences. \u003cem\u003eIn vitro\u003c/em\u003e transcription was done using the HiScribe\u0026trade; T7 Quick High Yield RNA Synthesis Kit. The individual transcription reactions were performed in 25 \u0026micro;L of total volume. This included 0.5 \u0026micro;L of 100 \u0026micro;M T7 forward primer, 1.5 \u0026micro;L of 100 \u0026micro;M crRNA-encoding DNA oligonucleotide, 1.25 \u0026micro;L of T7 RNA polymerase, 9.25 \u0026micro;L of 2x NTP buffer and 12.5 \u0026micro;L of nuclease-free ddH2O. The reactions were carried out for 2 h at 37\u0026deg;C. The obtained crRNAs were used in Cas13a activity assays without additional purification.\u003c/p\u003e\u003cp\u003e\u003cb\u003eCas13a activity assays for testing crRNA performance\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThe Cas13a activity assays detailed below have been developed and described in our previous publications and were used in this work with minor modifications.\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e,\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e To determine the efficacy of each crRNA, Cas13a nuclease activity assays were conducted using Cas13a enzyme from \u003cem\u003eL. wadei\u003c/em\u003e \u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e which was synthesized and purified by GenScript Biotech (Piscataway, NJ). Depending on the number of crRNA and RNA targets run at the same time the assay was conducted using an Echo 525 acoustic liquid handler (Beckman Coulter, Indianapolis, IN) using the Plate Reformat software provided by the manufacturer as described earlier \u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e or performed manually.\u003c/p\u003e\u003cp\u003eThe Cas13a enzyme was stored and diluted using the storage buffer (50 mM Tris-HCl, 600 mM NaCl, 5% Glycerol, 2 mM DTT, pH 7.5). Each nuclease activity assay was performed in 20 \u0026micro;L reaction that included 1 \u0026micro;L of 1 \u0026micro;M Cas13a, 1 \u0026micro;L of 2 \u0026micro;M RNase alert v.2 (from RNaseAlert\u0026trade; QC System v2, ThermoFisher), 17.2 \u0026micro;L of nuclease assay buffer (40 mM Tris-HCl, 60 mM NaCl, 6m M MgCl2, pH 7.3), 0.4 \u0026micro;L of crRNA (from unpurified transcription reaction) and 0.4 \u0026micro;L of 30 mM target RNA. For each crRNA a total of six reactions were set up, with three target negative reactions and three target positive wells. First, a master mix containing all reaction components except for the crRNA and target RNA were distributed to a 384 well assay plate. A total volume of 19.2 \u0026micro;L of the master-mix was transferred to each well. Next, 0.4 \u0026micro;L crRNAs were transferred to the wells containing the master-mix in such a way that each crRNA was added to 6 subsequent wells in the reaction plate. Finally, 0.4 \u0026micro;L of the target RNAs were added to three of the wells for each crRNA. The Cas13a reaction plates were spun briefly in a centrifuge at approximately 1500 x g to bring all the liquid to the bottom of the wells and remove air bubbles. Immediately after spinning, the reaction plates were sealed using MicroAmp sealers. The plates were incubated without shaking in SpectraMax M3 plate reader (Molecular Devices, San Jose, CA) at 37\u0026deg;C and fluorescence was read from the bottom of the wells every 5 minutes for 2 hours using excitation at 490 nm, emission at 520 nm with auto cutoff on and PMT gain set to \u0026ldquo;medium\u0026rdquo; and 6 flashes per read and carriage speed set to \u0026ldquo;normal\u0026rdquo;.\u003c/p\u003e\u003cp\u003eThe integrated background corrected final fluorescence values reflecting the Cas13a RNase activation for each of the crRNAs was calculated by subtracting the sum of averages of fluorescence measured for template negative samples over the course of the experiment (25 measurements) from sum of averages for template positive samples.\u003c/p\u003e\u003cp\u003eEach of the tested crRNAs was classified as positive or negative based on the corrected integrated fluorescent signal, i.e. background subtracted, relative to the signal obtained for the perfectly matching target (target with zero mismatches). The assay was considered positive when the signal was equal or higher to 20% of the reference and negative for signal below 20%.\u003c/p\u003e\u003cp\u003e\u003cb\u003eIntegrated modified SHERLOCK assay\u003c/b\u003e\u003c/p\u003e\u003cp\u003eDetection of DNA and RNA version of the targets was conducted using two step assay broadly based on SHERLOCK detection method\u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e. The assay was performed in two steps: step one, recombinase polymerase amplification (RPA) based target amplification combined with reverse transcription for RNA targets and step two: T7 based \u003cem\u003ein vitro\u003c/em\u003e transcription combined with Cas13a activity assay. The RPA amplification step was performed in 10.8\u0026ndash;11.3 \u0026micro;L total volume using TwistAmp\u0026reg; Liquid Basic kit (TwistDx Limited, Maidenhead, UK). The initial reaction mix was set up by combining 5 \u0026micro;L of the 2x TwistAmp reaction buffer, 1.8 \u0026micro;L of 10 mM dNTP mix (New England Biolabs), 1 \u0026micro;L TwistAmp 10x basic E mix, 0.5 \u0026micro;L of 10 \u0026micro;M of each amplification primer, 0.5 \u0026micro;L of Superscript III (ThermoFisher) and 0.5 \u0026micro;L of TwistAmp core reaction mix. In case of DNA templates, the Superscript III was omitted and replaced by nuclease-free water. The reagents were mixed by tube inversion or pipetting up and down. Next, 0.5 \u0026micro;L of magnesium acetate (MgOAc) and 1 \u0026micro;L of the template (target) DNA or RNA were added in separate drops to the lid of the tube. The tube was closed, and reagents mixed by inverting the tube 6 times. The reaction mixture was subsequently incubated at 37\u0026deg;C for one hour. The amplified targets were immediately used for detection in Cas13a assay (step two) or stored frozen at -80\u0026deg;C. The combined T7 based transcription and Cas13 detection was carried out in 30 \u0026micro;L reaction that included 1.25 \u0026micro;L of 1 \u0026micro;M Cas13a, 1.25 \u0026micro;L of 2 \u0026micro;M RNase alert v.2, 0.5 \u0026micro;L of unpurified crRNA rection mix, 0.25 \u0026micro;L of T7 RNA polymerase (from HiScribe\u0026trade; T7 Quick High Yield RNA Synthesis Kit), 5 \u0026micro;L of 2x NTP buffer (from HiScribe\u0026trade; T7 Quick High Yield RNA Synthesis Kit), 20.5 \u0026micro;L of nuclease assay buffer (40 mM Tris-HCl, 60 mM NaCl, 6m M MgCl\u003csub\u003e2\u003c/sub\u003e, pH 7.3) and 3 \u0026micro;L of the amplified target obtained in step one. For each detection six reaction were run including 3 negative controls (with TE buffer added in place of the amplified target) and 3 detection reactions. The reactions were incubated for 2 hours at 37\u0026deg;C in 384 reaction plates and read using the plate reader and standard Cas13a assay procedures as described above.\u003c/p\u003e\u003cp\u003e\u003cb\u003eDataset, data processing, and feature extraction\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThe data processing, and feature extraction methods described below have been previously published previously in our earlier work.\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e A dataset was constructed based on the results of the assays testing the performance of all the tested crRNAs with a panel of targets containing 13 coronavirus sequences representing SARS-CoV-2 and a several clades of nearest neighbor zoonotic coronaviruses. Each data entry included a list of positions of mismatches between the crRNA spacer and the corresponding target sequence together with a fluorescent signal obtained in the Cas13a activity assay using this crRNA spacer/target combination. To identify the mismatch positions the target sequences (converted to DNA sequence) and reverse complements of the crRNA spacers (also converted to DNA sequences) were compared. Mismatches were identified by applying binary labels for match/mismatch for each base and each spacer/target pairing. For each of the dataset entries 22 features were extracted or calculated from the mismatch data and target sequences as described previously\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e. The features most relevant for this work include: \u0026ldquo;n\u0026rdquo; \u0026ndash; the total number of mismatches between the crRNA spacer (guide sequence) and the target sequences, \u0026ldquo;mean\u0026rdquo; \u0026ndash; the mean value calculated for the spacer positions of all mismatches, \u0026ldquo;IQR\u0026rdquo; (interquartile range) \u0026ndash; the difference between Q1 and Q3 quartile values of all mismatch positions for the spacer sequence, \u0026ldquo;min\u0026rdquo; \u0026ndash; spacer position of a mismatch nearest to the crRNA hairpin (5\u0026rsquo; end of crRNA) and \u0026ldquo;max\u0026rdquo; - spacer position of a mismatch nearest to the 3\u0026rsquo; end of crRNA.\u003c/p\u003e\u003cp\u003eThe results of the Cas13a activity assays using a particular crRNA spacer/target combination were designated as either positive or negative based on the following criteria: a sample was evaluated as negative if the cumulative fluorescent, background subtracted, signal was less than 20% of the maximum signal obtained for the crRNA assay with target with no mismatches, and positive if it was greater than or equal to 20% of the maximum signal.\u003c/p\u003e\u003cp\u003e\u003cb\u003eModels and feature importance ranking\u003c/b\u003e\u003c/p\u003e\u003cp\u003eModels were built to classify the spacer/target combinations as producing positive or negative assay outcomes as described earlier\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e. Rule-based models such as RuleFit use groupings (ensembles) of linear models to build either classification or regression predictions that are comparable in accuracy as the best alternatives\u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e. However, their main advantage is in their interpretability, as each rule in the ensemble is a simple statement related to the individual features in the input dataset. This property of RuleFit allows for clear ranking of the relative importance of each feature and allows to better understand their data and the predictions.\u003c/p\u003e\u003cp\u003eThe classification model was generated in R using the Tidymodels series of packages\u003csup\u003e\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e. Rule based Learning Ensembles (RuleFit) were assembled with the XRF package \u003csup\u003e\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e. The number of trees contained in the ensemble was set to 2, maximum depth of the tree was set to 3, and the L1 regularization parameter was set to 0.01; all other parameters were set to defaults.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cb\u003eTesting of the crRNA performance.\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThe guides selected by the machine learning model were subsequently tested against the target analytes using a standard Cas13a activity assay with 1.5 nM final target concentration and unpurified in-vitro transcribed crRNAs in 20 \u0026micro;L total reaction volume as described in the methods section. We present the separate results focusing on each specific clade (i.e. color group), and the capability of each guide to maintain specificity within its group. A summary of the results is available in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e with the detailed results available in the supporting information.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cb\u003eBlack group: SARS-CoV-2/Wuhan-Hu-1 specific\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThe total of ten crRNAs specific for SARS-CoV-2 (black group) were selected for testing of their performance and specificity. Only guide sequences with zero mismatches to SARS-CoV-2 target and four or more mismatches to zoonotic targets were selected for this group. Each of the crRNAs was tested with SARS-CoV-2 and all 12 zoonotic targets (Detailed results available in Supplementary Figure \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e and Supplementary Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eAll ten crRNAs showed good activity for SARS-CoV-2 (Wuhan-Hu-1) target, however majority of them also showed activity with several of the nearest neighbor targets belonging to \u0026ldquo;dark red\u0026rdquo; target group. All the crRNAs except crRNA blk-4 and blk-7 were positive for BANAL-236 target and all but crRNAs blk-4, blk-6 and blk-7 were positive for RShSTT182 (which are the two closest neighbors tested). In addition, crRNAs blk-8 and blk-9 were positive for RaTG13 and crRNA blk-9 was also additionally positive for MP789.\u003c/p\u003e\u003cp\u003eThe only two crRNAs which were positive exclusively for SARS-CoV-2 (Wuhan-Hu-1) target were blk-4 and blk-7 (Supplementary Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003cb\u003eDark red group\u003c/b\u003e\u003c/p\u003e\u003cp\u003eFour crRNAs specific for \u0026ldquo;dark red\u0026rdquo; group of targets (BANAL-236, RaTG13, RShSTT182, MP789, and Guangxi-P4L) were tested: dkrd-1, dkrd-2, dkrd-3, and dkrd-5 (Detailed results available in Supplementary Figure S3 and Supplementary Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e). Due to a very high similarity of the targets from this group to SARS-CoV-2/Wuhan-Hu-1 the guide sequences for this crRNA group were designed using relaxed criteria without defining SARS-CoV-2/Wuhan-Hu-1 as the exclusion group (see Methods for details). As a result, all four of the tested crRNAs have between zero and two mismatches with SARS-CoV-2/Wuhan-Hu-1 target and were found positive for both \u0026ldquo;dark red\u0026rdquo; group of targets and SARS-CoV-2/Wuhan-Hu-1. Also, only the dkrd-1 and dkrd-2 crRNAs were found exclusively specific for \u0026ldquo;dark red\u0026rdquo; and SARS-CoV-2/Wuhan-Hu-1 targets. The crRNAs dkrd-3 and dkrd-5 were additionally found positive for targets belonging to the \u0026ldquo;yellow\u0026rdquo; group (SL-CoVZC45 and RpYN06).\u003c/p\u003e\u003cp\u003e\u003cb\u003eLight blue group\u003c/b\u003e\u003c/p\u003e\u003cp\u003eFive crRNAs specific for the light blue target (RsYN04) were tested: ltbl-21, ltbl-122, ltbl-127, ltbl-175, ltbl-177 (Detailed results available in Supplementary Figure S4 and Supplementary Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e). Only one crRNA (ltbl-177) was found to be completely specific and positive exclusively for the \u0026ldquo;light blue\u0026rdquo; target. The ltbl-21 was positive for the light blue target (RsYN04) and just above the threshold of positivity (22%) for RaTG13 (one of the \u0026ldquo;dark red\u0026rdquo; targets). The ltbl-122 crRNA failed to produce meaningful signal for any of the tested targets. One crRNA (ltbl-175) was positive for the \u0026ldquo;light blue\u0026rdquo; (RsYN04) and \u0026ldquo;dark blue\u0026rdquo; (RsYN03) targets and another one (ltbl-127) was found positive for the \u0026ldquo;light blue\u0026rdquo; (RsYN04) and two \u0026ldquo;yellow\u0026rdquo; (SL-CoVZC45 and RpYN06) targets.\u003c/p\u003e\u003cp\u003e\u003cb\u003eGreen group\u003c/b\u003e\u003c/p\u003e\u003cp\u003eFive crRNAs designed for detection of the \u0026ldquo;green\u0026rdquo; targets (BANAL-247 and RacCS203) were tested: grn-60, grn-87, grn-91, grn-97, and grn-133 (Detailed results available in Supplementary Figure S5 and Supplementary Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e). Two crRNAs (grn-87 and grn-97) were found to be positive exclusively for green targets. The other two crRNAs (grn-60 and grn-91) were found positive only for one of the \u0026ldquo;green\u0026rdquo; targets (RacCS203). Finally, the crRNA grn-133 was found positive for both of the \u0026ldquo;green\u0026rdquo; targets and at the threshold of positivity (20%) for the \u0026ldquo;dark blue\u0026rdquo; (RsYN03) target.\u003c/p\u003e\u003cp\u003e\u003cb\u003eDark blue group\u003c/b\u003e\u003c/p\u003e\u003cp\u003eFive crRNAs designed for detection of the \u0026ldquo;dark blue\u0026rdquo; target (RsYN03) were tested: dkbl-24, dkbl-25, dkbl-64, dkbl-116, and dkbl-150 (Detailed results available in Supplementary Figure S6 and Supplementary Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e). Only one of the tested crRNAs (dkbl-24) was found specific for the \u0026ldquo;dark blue\u0026rdquo; target. Two crRNAs (dkbl-64, dkbl-116) detected the \u0026ldquo;dark blue\u0026rdquo; target and one or both \u0026ldquo;yellow\u0026rdquo; targets (SL-CoVZC45 and RpYN06). The dkbl-25 crRNA detected one of the \u0026ldquo;dark red\u0026rdquo; targets (Guangxi-P4L) in addition to the \u0026ldquo;dark blue\u0026rdquo; target. The dkbl-25 was the least specific of the tested crRNA and generated signal above the positivity threshold for \u0026ldquo;dark blue\u0026rdquo; target, four \u0026ldquo;dark red\u0026rdquo; targets (RaTG13, RShSTT182, MP789, and Guangxi-P4L), both \u0026ldquo;yellow\u0026rdquo; targets and the \u0026ldquo;purple\u0026rdquo; target.\u003c/p\u003e\u003cp\u003e\u003cb\u003eYellow group\u003c/b\u003e\u003c/p\u003e\u003cp\u003eFive crRNAs designed for detection of the \u0026ldquo;yellow\u0026rdquo; targets (SL-CoVZC45 and RpYN06) were tested: ylw-45, ylw-75, ylw-83, ylw-86, and ylw-90 (Detailed results available in Supplementary Figure S7 and Supplementary Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e). While all the tested crRNAs gave positive signal with both \u0026ldquo;yellow\u0026rdquo; targets, only one of them (ylw-83) was specific to these targets only. Three crRNAs (ylw-45, ylw-75, and ylw-90) were additionally positive for the \u0026ldquo;dark blue\u0026rdquo; target. Two of these crRNAs (ylw-45 and ylw-90) were also additionally positive for some of the \u0026ldquo;dark red\u0026rdquo; targets: Guangxi-P4L in case of ylw-45 and MP789 in case of ylw-90. The ylw-75 crRNA was found to be positive for the \u0026ldquo;green\u0026rdquo; target (RacCS203) in addition to \u0026ldquo;dark blue\u0026rdquo; and \u0026ldquo;yellow\u0026rdquo;. Finally, the ylw-86 crRNA was positive for both \u0026ldquo;yellow\u0026rdquo; targets and one of the \u0026ldquo;dark red\u0026rdquo; targets (MP789).\u003c/p\u003e\u003cp\u003e\u003cb\u003ePurple group\u003c/b\u003e\u003c/p\u003e\u003cp\u003eFour crRNAs designed for detection of the \u0026ldquo;purple\u0026rdquo; target (Rco319) were tested: prpl-6, prpl-23, prpl-54, and prpl-105 (Detailed results available in Supplementary Figure S8 and Supplementary Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e). All of the tested crRNAs produced strong signal for the \u0026ldquo;purple target\u0026rdquo;, hover only one crRNA (prpl-6) was specific to purple target. In addition to \u0026ldquo;purple\u0026rdquo; the prpl-23 crRNA was positive for \u0026ldquo;dark blue\u0026rdquo; target, the prpl-54 crRNA was positive for one of the \u0026ldquo;dark red\u0026rdquo; targets (RShSTT182) and the prpl-105 crRNA was positive for one of the \u0026ldquo;yellow\u0026rdquo; targets (SL-CoVZC45).\u003c/p\u003e\u003cp\u003e\u003cb\u003ePerformance of the RuleFit based predictive model\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThe RuleFit algorithm-based predictive model was previously developed using experimental data to predict crRNA performance based on the number and distribution of the mismatches between guide and target sequences\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e. The RuleFit classifier provides an initial estimate for whether a guide-target pair will result in a positive signal, i.e. \u0026gt;20% of max signal. That algorithm was used as is to select initial guide sequences for this study. To design guides which would generate signal with the inclusion targets yet not to detect the exclusion group the threshold was set at 4 mismatches or greater with the exclusion group targets. Unexpectedly, the \u003cem\u003ein vitro\u003c/em\u003e testing results showed that multiple guide-target pairings with up to 6 mismatches could produce strong signal (e.g. 88% of max signal for SARS-CoV-2 specific guide black-9 with RshSTI182 target sequence).\u003c/p\u003e\u003cp\u003eThe model was retrained using the \u003cem\u003ein vitro\u003c/em\u003e data obtained in this study and performance of the modified model was assessed again. The confusion matrix (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e, panel A) compares the predictions of the updated model with the actual experimental data. The model shows an overall accuracy of 94%. Consistent with previous design rules, the feature with the overall highest global impact on the predictions was the total mismatch count (n). Increasing the threshold value to n\u0026thinsp;\u0026ge;\u0026thinsp;6 allow correct classification of 90% of the pairings. In contrast to the original findings where IQR was determined to be the second most important feature, after retraining the model, the \u0026ldquo;mean\u0026rdquo; feature replaced it as the second most important feature with IRQ moving to the third position (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e, panel B). The significance of the \u0026ldquo;mean\u0026rdquo; feature suggests that mismatch positions closer to the 3\u0026rsquo; end of the guide are less disruptive, which are in line with the earlier observations that shortening the guide sequence by up to seven nucleotides from the 3\u0026rsquo; end has little impact on crRNA performance\u003csup\u003e\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e. To determine the effect of shortening the guide sequence on the model predictions the model was retrained using a data set generated using just the first 20 positions of the guide/target pairing. As a result, it was found that for the shortened guides the threshold mismatch n value for 90% efficiency (Supplementary Figure S9, panel A) of predicting negative outcome dropped from 6 to 4 within the test set and the importance of the number mismatches for predictions increased significantly over other features (Supplementary Figure S9, panel B), with percentage of decisions explained by the number of mismatches increasing from 47\u0026ndash;70%.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cb\u003eLimit of detection (LOD) determination for SARS-CoV-2 DNA target for the integrated assay.\u003c/b\u003e\u003c/p\u003e\u003cp\u003eA series of seven 10-fold dilutions of the Wuhan-Hu-1 synthetic DNA target ranging from 250 pM to 250 aM and seven 10-fold dilutions of synthetic RNA target ranging from 1.5 nM to 1.5 fM were prepared and tested with the two step SHERLOCK type assay detailed in the methods. The target preparations were amplified in the first step of the assay using primer pair specifically designed for amplification of Wuhan-Hu-1 synthetic target. The crRNAs used for the Cas13a part of the assay were an equimolar mixture of two Wuhan-Hu-1 specific crRNAs black-4 and black-7. The negative controls (no target control \u0026ndash; NTC) was a Cas13a assay with TE buffer in place of the amplified target and positive control was a Cas13a assay with 1.5 nM Wuhan-Hu-1 RNA target.\u003c/p\u003e\u003cp\u003eThe results of the experiments are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e. The LOD for DNA targets was determined to be between 25 fM and 2.5 fM while the LOD using RNA targets was determined to be between 150 fM and 15 fM. The sensitivity of the assay for specific detection of Wuhan-Hu-1 synthetic target using a mixture of black-4 and black-7 crRNAs is approximately 6 times higher for DNA templates compared to RNA templates. The loss of sensitivity when using RNA templates seems to be the result of including of reverse transcription in the first step of the assay.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cb\u003eDetection of mixed RNA targets and performance of the assay for discrimination of SARS-CoV-2 and its zoonotic near neighbors.\u003c/b\u003e\u003c/p\u003e\u003cp\u003eWith the intent of testing the inhibitory effect of non-target RNAs inclusion in the sample matrix we conducted a mixed RNA pool test. Results of the experiment testing detection of RNA targets in the mixed sample are shown in the Supplementary Figure S10. The experiment was conducted using a sample containing mixture of all 13 synthetic targets tested in this study at 1.5 nM final mixed target concentration. A total of eleven Cas13 assays were conducted using selected individual crRNAs designed for detection of SARS-CoV-2 and 6 distinct clades of zoonotic viruses. The following crRNAs were used in this assay black-7 (specific for Wuhan-Hu-1 target), dkrd-1, dkrd-2 (specific for both Wuhan-Hu-1 and \u0026ldquo;dark red\u0026rdquo; zoonotic clade), dkbl-24 (specific for \u0026ldquo;dark blue\u0026rdquo; zoonotic clade), grn-87, grn-133 (mostly specific for \u0026ldquo;green\u0026rdquo; zoonotic clade), ltbl-127 (specific for \u0026ldquo;light blue\u0026rdquo; and \u0026ldquo;yellow\u0026rdquo; zoonotic clades), ltbl-177 (specific for \u0026ldquo;light blue\u0026rdquo; zoonotic clade), prpl-6 (specific for \u0026ldquo;purple\u0026rdquo; zoonotic clade), ylw-75 (specific for \u0026ldquo;yellow\u0026rdquo; and \u0026ldquo;dark blue\u0026rdquo; zoonotic clades) and ylw-86 (specific for \u0026ldquo;yellow\u0026rdquo; zoonotic clade and cross reacting with one of the \u0026ldquo;dark red\u0026rdquo; targets). All of the detection assays were positive based on the 20% maximum signal threshold with the lowest signal (30%) obtained for ylw-86 and highest (100%) for prpl-6 crRNA.\u003c/p\u003e\u003cp\u003e\u003cb\u003eDuplex assay for discrimination of SARS-CoV-2 and zoonotic near neighbors\u003c/b\u003e\u003c/p\u003e\u003cp\u003eAs a final proof of concept, we aimed to demonstrate the capability to distinguish between a specific target (SARS-CoV-2) and its near neighbors using a limited set of detection channels. The intent being the capability to utilize a simple assay to help distinguish known specific target from natural near-neighbors or engineered variants.\u003c/p\u003e\u003cp\u003eThe two-step modified SHERLOCK assay was conducted using the following RNA targets: Wuhan-Hu-1 (SARS-CoV-2) and seven targets representing all zoonotic near neighbor clades: BANAL-236 and Guangxi-P4L (dark red), RsYN04 (light blue), RacCS203 (green), RsYN03 (dark blue), RpYN06 (yellow) and Rc-o319 (purple). The synthetic RNA targets at two different concentrations were amplified using RT/RPA and each of the amplified targets were used for two T7/Cas13a detection assays one using a SARS-CoV-2 specific crRNA mixture (blk-4 and blk-7) and the other using zoonotic-near-neighbor-detection mixture containing 10 distinct crRNAs and designed to detect both SARS-CoV-2 and closely related sarbecoviruses (dkrd-1, dkrd-2, ltbl-127, ltbl-177, grn-87, grn-133, dkbl-24, ylw-75, ylw-86, prpl-6). The results of this assay are summarized in the Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e and Supplementary Figure S11. The results show that the SARS-CoV-2 specific reaction was positive for Wuhan-Hu-1 but also positive for its nearest zoonotic neighbor BANAL-236 but with much lower fluorescent signal. The other zoonotic targets were negative in this assay. For the assay using zoonotic-near-neighbor-detection crRNA mixture all targets were found positive with varying signal levels. Taking both assay into account it was possible to distinguish between three cases: (1) SARS-CoV-2 positive (strongly positive SARS-CoV-2 assay and positive zoonotic-near-neighbor-detection assay with significantly lower fluorescent signal), (2) BANAL-236 positive (both assays positive with comparable signal levels) and (3) zoonotic-near-neighbor positive (negative SARS-CoV-2 assay and positive zoonotic-near-neighbor-detection assay).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eThis study demonstrates the feasibility of developing detection assays capable of differentiation of targets of varying degree of variability with small number of detection channels using CRISPR/Cas based RNA/DNA target detection coupled with guide sequence optimization with machine learning model.\u003c/p\u003e\u003cp\u003eIn this work we used a model system in which the aim was to distinguish RNA target sequences representing SARS-CoV-2 (specific target) from a group of closely related sarbecoviruses. The example was chosen due to its relevance, yet the developmental pipeline could be applied to other viruses or systems of interest. The RuleFit based machine learning model which was previously trained on Lassa virus (LASV) sequences was used for initial selection of guide sequences for detection of the receptor binding domain of spike gene. This region of genome is known for relatively high level of divergence between SARS-CoV-2 and zoonotic near neighbors compared to the adjacent sequences. \u003cem\u003eIn vitro\u003c/em\u003e verification of the actual ranges of specificity of these guide sequences have shown that the model needed further optimization. We found out that the four mismatch threshold between guide and target as determined using LASV dataset is not sufficient to avoid cross-reactivity with targets belonging to the exclusion groups. Retraining the model using the new \u003cem\u003ein vitro\u003c/em\u003e data increasing the threshold to 6 mismatches allowed the model to achieve high level of accuracy. This final result highlights, the ubiquitous by now, knowledge that the quality of results from any machine learning model is dependent on the quality of the input data. While the LASV dataset was a good starting place to train the RuleFit model, use of the specific sarbecovirus training data further improved the results. Novel applications would therefore be able to use existing guide rules (e.g. LASV or sarbecovirus) for immediate results, while then retraining their model for subsequent optimization.\u003c/p\u003e\u003cp\u003eOur initial testing using modified two-step SHERLOCK assay taking advantage of simultaneous reverse transcription and RPA amplification followed by T7 based transcription and Cas13a detection shows that the assay has the sufficient sensitivity to detect the virus in concentrations present in the nasal wash during the acute phase of the COVID infection\u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e (LOD for RNA targets between 150 fM and 15 fM equivalent to target molecule concentrations between ~\u0026thinsp;2.5x10^4 and ~\u0026thinsp;2.5x10^3 copies/\u0026micro;L). The optimized guide sequence sets allowed us to build a two-channel detection assay which can not only detect SARS-CoV-2 but, at the same time, indicate the presence of one of the near-neighbor sarbecoviruses and distinguish situations in which both or just one of these (groups) of targets is present in the sample. This type of assays may be extremely useful for detection new zoonotic spillovers and use of genetically modified pathogens. While we would expect novel zoonotic, or even manmade, variants to be uncommon, our approach provides an early detection approach without massively expanding the assay requirements and costs. Upon a positive result for variants, the sample could be more thoroughly characterized for a more precise determination and the proper course of action.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003ch2\u003eCompeting Interests\u003c/h2\u003e\u003cp\u003eTAL and DAS are listed as inventors on a patent application for methods described herein. The remaining authors declare no competing interests.\u003c/p\u003e\u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e\u003cp\u003edeclaration\u003c/p\u003e\u003cp\u003eThis work was funded by the Defense Threat Reduction Agency (HDTRA1240013) and Office of Naval Research Base funding to the U.S. Naval Research Laboratory.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eDAS and TAL conceived the study, DAS, TAL and CMG designed the experimental protocols, TAL performed the experimental part of the study, DAS, TAL, SAD, ZTG and CMG \u0026ndash; analyzed the obtained experimental data, SND performed statistical and machine learning analysis of crRNA guide performance using RuleFit model, TAL wrote the first version of the manuscript. DAS, SAD, ZTG and CMG reviewed and made edits to the manuscript, TAL, ZTG and CMG prepared the manuscript figures, DAS obtained the funding for this work. All authors read and approved the submitted version of the manuscript.\u003c/p\u003e\u003ch2\u003eAcknowledgement\u003c/h2\u003e\u003cp\u003e This manuscript was approved for public release by DTRA and U.S. Naval Research Laboratory with unlimited distribution.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eAll data and code used in this study are available through the GitHub repository at https://github.com/NRL-CRISPR/CRISPR-rules.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eLi, H. Y. et al. A qualitative study of zoonotic risk factors among rural communities in southern China. \u003cem\u003eInt. Health\u003c/em\u003e. \u003cb\u003e12\u003c/b\u003e, 77\u0026ndash;85. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/inthealth/ihaa001\u003c/span\u003e\u003cspan address=\"10.1093/inthealth/ihaa001\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChomel, B. B., Belotto, A. \u0026amp; Meslin, F. X. Wildlife, exotic pets, and emerging zoonoses. \u003cem\u003eEmerg. Infect. Dis.\u003c/em\u003e \u003cb\u003e13\u003c/b\u003e, 6\u0026ndash;11. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3201/eid1301.060480\u003c/span\u003e\u003cspan address=\"10.3201/eid1301.060480\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2007).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eField, H. E. Bats and emerging zoonoses: henipaviruses and SARS. \u003cem\u003eZoonoses Public. Health\u003c/em\u003e. \u003cb\u003e56\u003c/b\u003e, 278\u0026ndash;284. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/j.1863-2378.2008.01218.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1863-2378.2008.01218.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2009).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMilne-Price, S., Miazgowicz, K. L. \u0026amp; Munster, V. J. The emergence of the Middle East respiratory syndrome coronavirus. \u003cem\u003ePathog Dis.\u003c/em\u003e \u003cb\u003e71\u003c/b\u003e, 121\u0026ndash;136. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/2049-632X.12166\u003c/span\u003e\u003cspan address=\"10.1111/2049-632X.12166\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2014).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCrits-Christoph, A. et al. Genetic tracing of market wildlife and viruses at the epicenter of the COVID-19 pandemic. \u003cem\u003eCell\u003c/em\u003e 187, 5468\u0026ndash;5482 e5411, (2024). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.cell.2024.08.010\u003c/span\u003e\u003cspan address=\"10.1016/j.cell.2024.08.010\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMozer, A. \u0026amp; Prost, S. An introduction to illegal wildlife trade and its effects on biodiversity and society. \u003cem\u003eForensic Sci. International: Anim. Environ.\u003c/em\u003e \u003cb\u003e3\u003c/b\u003e, 100064. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.fsiae.2023.100064\u003c/span\u003e\u003cspan address=\"10.1016/j.fsiae.2023.100064\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHughes, A. et al. Determining the sustainability of legal wildlife trade. \u003cem\u003eJ. Environ. Manage.\u003c/em\u003e \u003cb\u003e341\u003c/b\u003e, 117987. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.jenvman.2023.117987\u003c/span\u003e\u003cspan address=\"10.1016/j.jenvman.2023.117987\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAhteensuu, M. S. \u0026amp; Biology Genome Editing, and the Risk of Bioterrorism. \u003cem\u003eSci. Eng. Ethics\u003c/em\u003e. \u003cb\u003e23\u003c/b\u003e, 1541\u0026ndash;1561. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s11948-016-9868-9\u003c/span\u003e\u003cspan address=\"10.1007/s11948-016-9868-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2017).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSabra, D. M., Krin, A., Romeral, A. B., Friess, J. L. \u0026amp; Jeremias, G. Anthrax revisited: how assessing the unpredictable can improve biosecurity. \u003cem\u003eFront. Bioeng. Biotechnol.\u003c/em\u003e \u003cb\u003e11\u003c/b\u003e, 1215773. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fbioe.2023.1215773\u003c/span\u003e\u003cspan address=\"10.3389/fbioe.2023.1215773\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGerstein, D. M., Espinosa, B. \u0026amp; Leidy, E. \u003cem\u003eN. Emerging technology and risk analysis - synthetic pandemics\u003c/em\u003e (Homeland Security Operational Analysis Center operated by the RAND Corporation, 2024).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKamalrathne, T., Amaratunga, D., Haigh, R., Kodituwakku, L. \u0026amp; Rupasinghe, C. Epidemic and Pandemic Preparedness and Response in a Multi-Hazard Context: COVID-19 Pandemic as a Point of Reference. \u003cem\u003eInt. J. Environ. Res. Public. Health\u003c/em\u003e. \u003cb\u003e21\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/ijerph21091238\u003c/span\u003e\u003cspan address=\"10.3390/ijerph21091238\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLiu, Q. et al. Advances in the application of molecular diagnostic techniques for the detection of infectious disease pathogens (Review). \u003cem\u003eMol. Med. Rep.\u003c/em\u003e \u003cb\u003e27\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3892/mmr.2023.12991\u003c/span\u003e\u003cspan address=\"10.3892/mmr.2023.12991\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePatchsung, M. et al. Clinical validation of a Cas13-based assay for the detection of SARS-CoV-2 RNA. \u003cem\u003eNat. Biomed. Eng.\u003c/em\u003e \u003cb\u003e4\u003c/b\u003e, 1140\u0026ndash;1149. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41551-020-00603-x\u003c/span\u003e\u003cspan address=\"10.1038/s41551-020-00603-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGreen, C. M. et al. Quantum Dot-Based Molecular Beacons for Quantitative Detection of Nucleic Acids with CRISPR/Cas(N) Nucleases. \u003cem\u003eACS Nano\u003c/em\u003e. \u003cb\u003e16\u003c/b\u003e, 20693\u0026ndash;20704. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1021/acsnano.2c07749\u003c/span\u003e\u003cspan address=\"10.1021/acsnano.2c07749\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGootenberg, J. S. et al. Multiplexed and portable nucleic acid detection platform with Cas13, Cas12a, and Csm6. \u003cem\u003eScience\u003c/em\u003e \u003cb\u003e360\u003c/b\u003e, 439\u0026ndash;444. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1126/science.aaq0179\u003c/span\u003e\u003cspan address=\"10.1126/science.aaq0179\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2018).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGuo, X. et al. Transcriptome-wide Cas13 guide RNA design for model organisms and viral RNA pathogens. \u003cem\u003eCell. Genomics\u003c/em\u003e. \u003cb\u003e1\u003c/b\u003e, 100001. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.xgen.2021.100001\u003c/span\u003e\u003cspan address=\"10.1016/j.xgen.2021.100001\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2021).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWessels, H. H. et al. Massively parallel Cas13 screens reveal principles for guide RNA design. \u003cem\u003eNat. Biotechnol.\u003c/em\u003e \u003cb\u003e38\u003c/b\u003e, 722\u0026ndash;727. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41587-020-0456-9\u003c/span\u003e\u003cspan address=\"10.1038/s41587-020-0456-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMolina Vargas, A. M. et al. New design strategies for ultra-specific CRISPR-Cas13a-based RNA detection with single-nucleotide mismatch sensitivity. \u003cem\u003eNucleic Acids Res.\u003c/em\u003e \u003cb\u003e52\u003c/b\u003e, 921\u0026ndash;939. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/nar/gkad1132\u003c/span\u003e\u003cspan address=\"10.1093/nar/gkad1132\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDur\u0026aacute;n-Vinet, B. et al. CRISPR-Based Environmental Biosurveillance Assisted via Artificial Intelligence Design of Guide-RNAs. \u003cem\u003eEnviron. DNA\u003c/em\u003e. \u003cb\u003e7\u003c/b\u003e, e70105. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/edn3.70105\u003c/span\u003e\u003cspan address=\"10.1002/edn3.70105\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLeski, T. A. et al. Machine learning for design of degenerate Cas13a crRNAs using lassa virus as a model of highly variable RNA target. \u003cem\u003eSci. Rep.\u003c/em\u003e \u003cb\u003e13\u003c/b\u003e, 6506. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41598-023-33494-4\u003c/span\u003e\u003cspan address=\"10.1038/s41598-023-33494-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTemmam, S. et al. Bat coronaviruses related to SARS-CoV-2 and infectious for human cells. \u003cem\u003eNature\u003c/em\u003e \u003cb\u003e604\u003c/b\u003e, 330\u0026ndash;336. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41586-022-04532-4\u003c/span\u003e\u003cspan address=\"10.1038/s41586-022-04532-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKatoh, K. \u0026amp; Standley, D. M. MAFFT multiple sequence alignment software version 7: improvements in performance and usability. \u003cem\u003eMol. Biol. Evol.\u003c/em\u003e \u003cb\u003e30\u003c/b\u003e, 772\u0026ndash;780. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/molbev/mst010\u003c/span\u003e\u003cspan address=\"10.1093/molbev/mst010\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2013).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePersson, S., Larsson, C., Simonsson, M. \u0026amp; Ellstrom, P. rprimer: an R/bioconductor package for design of degenerate oligos for sequence variable viruses. \u003cem\u003eBMC Bioinform.\u003c/em\u003e \u003cb\u003e23\u003c/b\u003e, 239. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12859-022-04781-0\u003c/span\u003e\u003cspan address=\"10.1186/s12859-022-04781-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSchultzhaus, Z., Wang, Z. \u0026amp; Stenger, D. Systematic analysis, identification, and use of CRISPR/Cas13a-associated crRNAs for sensitive and specific detection of the \u003cem\u003elcrV\u003c/em\u003e gene of \u003cem\u003eYersinia pestis\u003c/em\u003e. \u003cem\u003eDiagn. Microbiol. Infect. Dis.\u003c/em\u003e \u003cb\u003e99\u003c/b\u003e, 115275. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.diagmicrobio.2020.115275\u003c/span\u003e\u003cspan address=\"10.1016/j.diagmicrobio.2020.115275\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2021).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSpangler, J. R., Leski, T. A., Schultzhaus, Z., Wang, Z. \u0026amp; Stenger, D. A. Large scale screening of CRISPR guide RNAs using an optimized high throughput robotics system. \u003cem\u003eSci. Rep.\u003c/em\u003e \u003cb\u003e12\u003c/b\u003e, 13953. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41598-022-17474-8\u003c/span\u003e\u003cspan address=\"10.1038/s41598-022-17474-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGootenberg, J. S. et al. Nucleic acid detection with CRISPR-Cas13a/C2c2. \u003cem\u003eScience\u003c/em\u003e \u003cb\u003e356\u003c/b\u003e, 438\u0026ndash;442. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1126/science.aam9321\u003c/span\u003e\u003cspan address=\"10.1126/science.aam9321\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2017).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKellner, M. J., Koob, J. G., Gootenberg, J. S., Abudayyeh, O. O. \u0026amp; Zhang, F. SHERLOCK: nucleic acid detection with CRISPR nucleases. \u003cem\u003eNat. Protoc.\u003c/em\u003e \u003cb\u003e14\u003c/b\u003e, 2986\u0026ndash;3012. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41596-019-0210-2\u003c/span\u003e\u003cspan address=\"10.1038/s41596-019-0210-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2019).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFriedman, J. H. \u0026amp; Popescu, B. A. Predictive learning via rule ensembles. \u003cem\u003eAnnals Appl. Stat.\u003c/em\u003e \u003cb\u003e2\u003c/b\u003e, 916\u0026ndash;954 (2008).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKuhn, M. \u0026amp; Wickham, H. \u003cem\u003eTidymodels: a collection of packages for modeling and machine learning using tidyverse principles.\u003c/em\u003e, (2020). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.tidymodels.org\u003c/span\u003e\u003cspan address=\"https://www.tidymodels.org\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFriedman, J. H. \u0026amp; Popescu, B. E. Predictive learning via rule ensembles. \u003cem\u003eannals Appl. Stat.\u003c/em\u003e \u003cb\u003e2\u003c/b\u003e, 916\u0026ndash;954 (2008).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAbudayyeh, O. O. et al. RNA targeting with CRISPR-Cas13. \u003cem\u003eNature\u003c/em\u003e \u003cb\u003e550\u003c/b\u003e, 280\u0026ndash;284. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/nature24049\u003c/span\u003e\u003cspan address=\"10.1038/nature24049\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2017).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSavela, E. S. et al. Quantitative SARS-CoV-2 Viral-Load Curves in Paired Saliva Samples and Nasal Swabs Inform Appropriate Respiratory Sampling Site and Analytical Test Sensitivity Required for Earliest Viral Detection. \u003cem\u003eJ. Clin. Microbiol.\u003c/em\u003e \u003cb\u003e60\u003c/b\u003e, e0178521. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1128/JCM.01785-21\u003c/span\u003e\u003cspan address=\"10.1128/JCM.01785-21\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-7041916/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7041916/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eWe have developed a CRISPR/Cas based assay able to distinguish between two ranges of closely related RNA targets using two detection channels. This required a pipeline to design RNA guide sets with the right degree of specificity. We tested our approach using SARS-CoV-2 and zoonotic near-neighbor sarbecoviruses. Using pre-existing guide design rules, we utilized a machine learning based model to design and optimize guide sets for specific detection of SARS-CoV-2 and separately to its nearest neighbors. The \u003cem\u003ein vitro\u003c/em\u003e testing of the guide sequences has shown that Cas13 assays can tolerate more mismatches than assumed based on previous guide design rules. Mismatches located closer to the 3\u0026rsquo; end of the guide and mismatches evenly distributed throughout the guide resulted in a smaller impact on the guide\u0026rsquo;s ability to activate the Cas enzyme. Modified SHERLOCK assay for detection and discrimination of SARS-CoV-2 and its zoonotic coronaviruses was developed using optimized sets of guides. The final assay was able to classify the targets into three classes 1) SARS-Co-V2, 2) closest known SARS-Co-V2 near-neighbor BANAL-236 and 3) the remaining zoonotic near-neighbors. This approach provides value through early detection of natural and engineered variants.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e","manuscriptTitle":"Discrimination of ranges of closely related RNA targets using CRISPR based detection assay developed using machine learning based optimization","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-08-08 21:32:48","doi":"10.21203/rs.3.rs-7041916/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-11-24T08:59:09+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-11-21T00:12:25+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-11-08T20:45:55+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"285084802549699760280271914106470723771","date":"2025-10-28T10:00:02+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"51927927182173282342395623135744575953","date":"2025-10-22T15:16:41+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-08-15T17:47:12+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"157033526838948686042302002938823633228","date":"2025-08-05T15:51:05+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-08-03T00:25:36+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-07-23T17:16:20+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-07-11T19:53:14+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-07-09T16:55:35+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2025-07-09T16:52:28+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"6cf72290-2785-4510-b291-eac00ba8e4c1","owner":[],"postedDate":"August 8th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"in-revision","subjectAreas":[{"id":52693663,"name":"Biological sciences/Biological techniques"},{"id":52693664,"name":"Biological sciences/Biotechnology"},{"id":52693665,"name":"Biological sciences/Computational biology and bioinformatics"},{"id":52693666,"name":"Biological sciences/Microbiology"}],"tags":[],"updatedAt":"2026-04-06T07:26:29+00:00","versionOfRecord":[],"versionCreatedAt":"2025-08-08 21:32:48","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7041916","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7041916","identity":"rs-7041916","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}