CriSNPr, a single interface for the curated and de novo design of gRNAs for CRISPR diagnostics using diverse Cas systems

  1. Asgar H Ansari
  2. Manoj Kumar
  3. Sajal Sarkar
  4. Souvik Maiti
  5. Debojyoti Chakraborty

  • Curated by eLife

    eLife logo

    Evaluation Summary:

    The web-based software developed in this study will be of interest to researchers who develop CRISPR-based diagnostic methods. The use of CRISPR-Cas to rapidly identify specific mutations in both cancer and infection is an evolving field with good potential to play a role in future research and diagnostics. This software will facilitate the implementation of such technologies and is therefore useful.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 agreed to share their name with the authors.)

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

CRISPR-based diagnostics (CRISPRDx) have improved clinical decision-making, especially during the COVID-19 pandemic, by detecting nucleic acids and identifying variants. This has been accelerated by the discovery of new and engineered CRISPR effectors, which have expanded the portfolio of diagnostic applications to include a broad range of pathogenic and non-pathogenic conditions. However, each diagnostic CRISPR pipeline necessitates customized detection schemes based on the fundamental principles of the Cas protein used, its guide RNA (gRNA) design parameters, and the assay readout. This is especially relevant for variant detection, a low-cost alternative to sequencing-based approaches for which no in silico pipeline for the ready-to-use design of CRISPRDx currently exists. In this manuscript, we fill this lacuna using a unified web server, CriSNPr (CRISPR-based SNP recognition), which provides the user with the opportunity to de novo design gRNAs based on six CRISPRDx proteins of choice ( Fn /en Fn Cas9, Lw Cas13a, Lb Cas12a, Aa Cas12b, and Cas14a) and query for ready-to-use oligonucleotide sequences for validation on relevant samples. Furthermore, we provide a database of curated pre-designed gRNAs as well as target/off-target for all human and SARS-CoV-2 variants reported thus far. CriSNPr has been validated on multiple Cas proteins, demonstrating its broad and immediate applicability across multiple detection platforms. CriSNPr can be found at http://crisnpr.igib.res.in/ .

Article activity feed

  1. Version published to 10.7554/elife.77976 on eLife
  2. eLife

    Author Response

    Reviewer #2 (Public review):

    Ansari et al. describe a web-based software for the design of guide RNA (gRNA) sequences and primers for CRISPR-Cas-based identification of single nucleotide variants (SNVs). The use of CRISPR-Cas to rapidly identify specific mutations in both cancer and infection is an evolving field with good potential to play a role in future research and diagnostics.

    The software described by Ansari et al. is easy to use for the design of guide RNAs. The most important question is how good the gRNAs that the software suggests are. As such, the manuscript would benefit from better describing the parameters used for the gRNA design as well as including more validation experiments. Clearly, the scope of the manuscript is not about developing different detection methods, but I would argue that performing more wet lab experiments is needed to support the usability of the software.

    We thank the reviewer for taking interest in this manuscript and raising an important point about increasing the number of targets for our wet lab experiments. To address this, we have tried to include more supporting data in the updated version of the manuscript.

    Reviewer #3 (Public review):

    This manuscript by Ansari and coworkers describes CriSNPr, a tool for designing gRNAs for CRISPR-based diagnostics for SNP detection. CriSNPr allows one to design assays to detect human and SARS-CoV-2 mutations, positioning the mismatches for optimal detection based on results from the literature. Designs can be generated for six different CRISPR effector proteins. The authors test their approach by designing assays to detect a single SNV using three different CRISPR effectors. A strength of the manuscript is that the method does appear to work, at least for the E484K mutation, for multiple CRISPR effector proteins.

    The weaknesses of this manuscript are the lack of data demonstrating that the method works. There is only one very small experimental demonstration using a single mutation (Figure 4) and some very high-level analyses using two SNP/SNV databases (Figure 5). The authors do not provide any data to answer any basic questions about how well their designs work, how fast and easy it is to run their method, or which designs are predicted to work better than others. These weaknesses ultimately limit the impact of the work on the field, as it is not clear what the benefits of using the author's approach are versus simply applying the rules for the individual CRISPR effector proteins outlined in Figure 1 of the manuscript.

    We thank the reviewer for taking interest in this manuscript and appreciate the constructive feedback and suggestions. In the new version of this paper, we've added more data to back up other SNVs with different CRISPR systems and the CriSNPr pipeline for sgRNA design. Even in these datasets, we see that for particular SNVs, the choice of the CRISPR system used might affect the sensitivity of detecting the mutation (Figures 5 and 6). This would be a huge task to do again for multiple targets and targeting systems, which is outside the scope of this study. Importantly, such large datasets are currently missing for the different CRISPRDx systems since we have not come across studies where users have comparatively determined the best methodology for their assay. In our opinion, criSNPr gives users this opportunity by providing a unified platform, and our validation assays show how this can be done in a relatively fast manner.

    A stand-alone version of the server is made available for download at https://github.com/asgarhussain/CriSNPr to increase its speed and accessibility for the end user.

    Addressing the point of determining which crRNAs work best for a given assay requires a large amount of data on target SNPs for individual Cas systems, which is currently scarce. In the current version of CriSNPr, we have considered prioritizing crRNA mismatch-sensitive positions based on original published studies. For example, for AaCas12b, mismatch positions are ranked as follows: 1&4 > 1&5 > 4&11 > 4&16 > 5&8 > 5&11 > 16&19. Similarly, crRNA mismatch-sensitive positions for individual Cas systems (as shown in Figure 1) have been used to prioritize crRNAs. Improving on these design principles further would require studying the biology of individual Cas:DNA/RNA interactions, which is beyond the scope of this study. However, in the updated version of the CriSNPr, we attempted to improve the scoring algorithm by taking into account off-targets for a crRNA design, and priority is given to the combinatorial positions with the fewest off-targets as well as the weightage of their efficacy.

  3. eLife

    Evaluation Summary:

    The web-based software developed in this study will be of interest to researchers who develop CRISPR-based diagnostic methods. The use of CRISPR-Cas to rapidly identify specific mutations in both cancer and infection is an evolving field with good potential to play a role in future research and diagnostics. This software will facilitate the implementation of such technologies and is therefore useful.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 agreed to share their name with the authors.)

  4. eLife

    Reviewer #1 (Public Review):

    In this study, Ansari et al. have created a web platform called CriSNPr which serves two purposes:

    1. It provides a set of pre-designed CRISPR RNAs for dbSNP-annotated Single Nucleotide Variants (SNVs) in the human genome and variants of concern in the SARS-CoV-2 genome.

    2. For unannotated/novel SNVs in either the human or SARS-CoV-2 genomes, it designs CRISPR RNAs de novo, based on sequence information provided by the user.
    For both options, the platform focuses on six different CRISPR/Cas systems currently in use for CRISPR/Cas-based diagnostics, five of which - Fn/enFnCas9, LbCas12a, AaCas12b, and Cas14a - are DNA-targeting, and one of which, LwCas13a, is RNA-targeting. In addition to CRISPR RNA design, CriSNPr also identifies PCR primer pairs that could be used to generate amplicons for downstream testing and validation. Overall, the authors have clearly defined the "back-end" strategy of CriSNPr and the mismatch criteria that were considered for each CRISPR/Cas system for CRISPR RNA design. They also provide information about the proportion of dbSNP-annotated SNVs that can be targeted by each CRISPR/Cas systems, with Cas14a and LwCas13a being the most versatile and widely targeting. Lastly, the authors have experimentally demonstrated the utility of CriSNPr for designing reagents to detect a specific SARS-CoV-2 variant, S gene containing E484K mutation, using FnCas9, AaCas12b and Cas14a. However, the design of CRISPR RNAs and PCR primers for the detection of SNVs in human genomic DNA was not demonstrated. Given the size, complexity and diploid nature of the human genome, this would have enhanced the significance of the study.

    CRISPR/Cas-based diagnostics have shown great promise for sensitive/low-cost detection of nucleic acids of infectious agents as well as genetic mutations in humans, including SNVs. Some of the major bottlenecks for CRISPR/Cas-based SNV detection include: (i) the choice of CRISPR/Cas system and (ii) the fast and accurate design of specific CRISPR RNAs and PCR primers. In CriSNPr, by including six different CRISPR/Cas systems and by generating PCR primers, the authors fill an important lacuna and provide a rapid and easy, yet adaptable, platform for developing diagnostic workflows. This is the major strength of the study and I foresee that this platform will greatly accelerate the field of CRISPR/Cas-based diagnostics for SNV detection. At the same time, there is room for further development, as the authors point out - the inclusion of other organisms for which SNV information is readily available and linked to distinct phenotypes, the estimation of CRISPR RNA sensitivity and specificity for all six systems, the enhancement of the platform with additional CRISPR/Cas systems as and when they are developed for diagnostics.

  5. eLife

    Reviewer #2 (Public Review):

    Ansari et al. describes a web-based software for the design of guide RNA (gRNAs) sequences and primers for CRISPR-Cas-based identification of single nucleotide variants (SNVs). The use of CRISPR-Cas to rapidly identify specific mutations in both cancer and infection is an evolving field with good potential to play a role in future research and diagnostics.

    The software described by Ansari et al. is easy to use for the design of guide RNAs. The most important question is how good the gRNAs that the software suggests are. As such, the manuscript would benefit from better describing parameter used for the gRNA design, as well as including more validation experiments. Clearly, the scope of the manuscript is not about developing the different detection methods, but I would argue that performing more wet lab experiments is needed to support the usability of the software.

  6. eLife

    Reviewer #3 (Public Review):

    This manuscript by Ansari and coworkers describes CriSNPr, a tool for designing gRNAs for CRISPR-based diagnostics for SNP detection. CriSNPr allows one to design assays to detect human and SARS-CoV-2 mutations, positioning the mismatches for optimal detection based on results from the literature. Designs can be generated for 6 different CRISPR effector proteins. The authors test their approach by designing assays to detect a single SNV using three different CRISPR effectors. A strength of the manuscript is that the method does appear to work, at least for the E484K mutation, for multiple CRISPR effector proteins.

    The weaknesses of this manuscript are the lack of data demonstrating that the method works. There is only one very small experimental demonstration using a single mutation (Figure 4), and some very high-level analysis using two SNP/SNV databases (Figure 5). The authors do not provide any data to answer any basic questions about how well their designs work, how fast and easy it is to run their method, or which designs are predicted to work better than others. These weaknesses ultimately limit the impact of the work on the field, as it is not clear what the benefits of using the author's approach are versus simply applying the rules for the individual CRISPR effector proteins outlined in Figure 1 of the manuscript.

  7. Version published to 10.1101/2022.02.17.479653v1 on bioRxiv