Precise in-field molecular diagnostics of crop diseases by smartphone-based mutation-resolved pathogenic RNA analysis

The study addresses the need for simple, low-cost, in-field molecular diagnostics to manage crop diseases, especially fungal pathogens that threaten global food security and often drive overuse of pesticides. Existing phenotypic and imaging methods can detect symptoms but typically cannot identify the pathogen species, virulence, or drug resistance. While nucleic acid-based diagnostics can, in principle, detect any pathogen and reveal genetic determinants of virulence and resistance, they are complex and largely confined to laboratories due to reliance on nucleic acid amplification. The authors propose an amplification-free, smartphone-integrated genetic assay capable of providing pathogen identity, viability, and resistance information directly in the field to enable early detection and precision disease management.

The paper reviews conventional phenotypic approaches, including symptom observation and foliage imaging (hyperspectral, thermographic, infrared thermometric), which enable non-invasive in-field measurements but generally lack pathogen specificity. Genotypic and immunological methods provide pathogen information, with nucleic acid assays offering broad coverage of pathogen species and insights into virulence and drug resistance. However, current nucleic acid methods typically require amplification and laboratory infrastructure, limiting field application. The authors also reference the importance of genetic markers (e.g., CYP51 mutations) for fungicide resistance, and note existing high-throughput sequencing-based tools (e.g., MARPLE) that, while powerful, remain complex, require amplification and specialized instruments, and are costly for routine field use.

Assay principle: The assay detects fungal RNA via a toehold-mediated strand displacement (TMSD) reaction using a double-stranded DNA probe (DProbe) composed of a cis and a trans strand with forward and reverse toeholds. The reverse toehold includes a C–Ag(I)–C artificial base pair. Target fungal RNA hybridizes to the forward toehold, initiating strand displacement that releases the trans strand and Ag(I) ions. Released Ag(I) inhibits urease activity, decreasing ammonia production from urea, which is visualized by the pH indicator phenol red (yellow at pH < 6.8; red at pH > 8.2). Thus, target RNA presence is transduced into a colorimetric readout. Probe design and optimization: Short RNAs (22 nt) on the internal transcribed spacer (ITS) of Puccinia striiformis (Pst) were screened to identify optimal binding sites. Lengthening the forward toehold (from 8 to 15 nt) improved detection of longer fungal RNAs. For single-nucleotide mutation (SNM) detection (CYP51 Y134F), DProbes with a fixed 7-nt reverse toehold and 5–25 nt forward toeholds were tested using fluorescence and electrophoresis to maximize discrimination between mutant (F134) and wild-type (Y134) RNAs; a 17-nt forward toehold provided the best discrimination. Assay implementation in solution: Typical reaction used DProbe (e.g., 100 nM) incubated with sample RNA for ~10 min at room temperature, followed by addition of urease (1 nM), urea (500 mM), and phenol red (250 µM). Absorbance at 560 nm was measured; visual color change was photographed. Electrophoresis and fluorescence validated TMSD. Origami paper device: Wax printing created hydrophobic barriers for a foldable paper (“origami”) device with sequential reagent zones (DProbe, enzyme, indicator). Folding initiated TMSD then color reporting. Paper substrate pore size and surfactant modification improved sensitivity and uniformity. A smartphone image-processing algorithm located detection spots and computed a green pixel ratio (GPR) to quantify color changes. Devices included eight loading sites enabling multiplexing for six wheat fungi (Pst, Pgt, Pt, Bgt, Fg, Rc) plus controls. Sample preparation: Standard TRIzol-based extraction was used in lab validations for fungi and plant tissues. For field use, two rapid extraction strategies were developed: (1) quartz sand grinding with simple buffer (fast but leaf pigments could affect color quantification), and (2) a polyvinyl alcohol microneedle (MN) patch (11×11 array, 600 µm needle height) pressed into leaves for 10 s, then eluted in 50 µL water, providing high-purity nucleic acids in ~1 min without instruments. Smartphone integration: A 3D-printed attachment enabled contamination-minimized, pull-based operation and standardized imaging. A mobile app guided users, analyzed images (GPR), and provided diagnostic results and treatment recommendations. Validation studies: Sensitivity and specificity were evaluated for Pst in solution and on paper; cross-reactivity was tested among congeneric rust fungi (Pst, Pgt, Pt) and with other wheat pathogens (Bgt, Fg, Rc). Field-collected wheat samples (n=32) were tested by both the paper assay and qPCR to estimate concordance. Early infection detection was assessed by time-course sampling post-inoculation (0–14 days) with both absorbance and paper-based GPR readouts, compared against symptom observation and qPCR. Viability discrimination was assessed using mixtures of viable/dead Pst spores (0–100% viable) with both colorimetric assay and qPCR. Mutation detection was validated via dilution of Y134F mutant in wild-type background (down to 0.1%), field isolate testing (n=8) with sequencing confirmation, and fungicide response (EC50 to triadimefon) assays. Key reagents and conditions: Example concentrations included DProbe 100–400 nM, fungal RNA up to 600 nM for TMSD validation, urease 1–10 nM, urea 500 mM, phenol red 250 µM. TMSD reaction time ~10 min at room temperature; total sample-to-answer time ~10 min with MN extraction in-field.

• Amplification-free detection: The assay detected Puccinia striiformis (Pst) RNA at 1.0 ng/µL without nucleic acid amplification (Welch’s t test: P < 0.05), with comparable sensitivity on paper to solution.
• Multiplexing and specificity: Six wheat pathogens (Pst, Pgt, Pt, Bgt, Fg, Rc) were visually and specifically detected on a single origami paper without cross-interference. Congeneric rust fungi were discriminated using species-specific ITS-targeting DProbes.
• Field sample performance: In 32 field-collected wheat leaves, the paper assay showed 95.3% positive prediction and 91.0% negative prediction versus qPCR for Pst infection.
• Early detection: Infection by Pst was detected as early as 3 days post inoculation by the colorimetric assay (absorbance drop significant at day 3; P < 0.05) and by paper (GPR; P < 0.01), preceding visible symptoms (day 10) by 7 days and matching qPCR early-detection timing.
• Viability discrimination: The colorimetric signal decreased with increasing proportion of viable Pst spores (0–100%), both with and without wheat RNA matrix, indicating detection of viable pathogens via RNA targeting; qPCR showed no significant difference across mixtures, underscoring its inability to distinguish viable from dead cells.
• Mutation detection and resistance inference: Optimized DProbe (17-nt forward toehold) achieved single-nucleotide discrimination of CYP51 Y134F and detected as low as 0.1% mutant in 99.9% wild-type background. In eight field isolates, four were mutation-free and four carried heterozygous Y134F, confirmed by sequencing; isolates with Y134F exhibited substantially higher EC50 to triadimefon than wild-type isolates, indicating resistance.
• In-field workflow: With MN-based extraction (~1 min), origami paper assay, and smartphone app, the sample-to-result process was completed in ~10 min. Estimated cost was ~$0.30 per test. The app provided user guidance and treatment recommendations.
• Broader applicability: Beyond fungi, proof-of-concept detection of a plant pathogenic bacterium (Pseudomonas syringae) and a virus (barley stripe mosaic virus) demonstrated potential breadth across pathogen types.

The findings demonstrate that an amplification-free, smartphone-readable nucleic acid assay can achieve sensitive, specific, and multiplexed detection of crop pathogens directly in the field. By exploiting TMSD coupled to Ag(I)-mediated urease inhibition, target RNA recognition is converted into a robust colorimetric signal. The platform addresses key gaps in current field diagnostics by providing: (1) early detection at the pre-symptomatic stage comparable to qPCR, (2) pathogen identification at species level to guide treatment choice, (3) discrimination between viable and dead pathogens to better estimate disease risk, and (4) single-nucleotide mutation detection to inform fungicide resistance and precise management strategies. Integration with rapid MN-based nucleic acid extraction and a smartphone app enables a practical sample-to-answer workflow suitable for non-expert users. Compared to imaging-based phenotyping, the assay adds genotypic specificity and resistance information; compared to lab-based nucleic acid methods, it removes the need for amplification and instruments, reducing complexity and cost while enabling in-field deployment.

The study introduces a low-cost (~$0.30), rapid (~10 min), amplification-free, and smartphone-integrated paper assay for in-field molecular diagnostics of crop diseases. It enables multiplexed detection of major wheat pathogens, early infection identification, viable pathogen discrimination, and mutation-resolved detection of fungicide resistance (e.g., CYP51 Y134F). The platform’s programmability allows expansion to other pathogens, including bacteria and viruses, advancing precision crop disease management. Future work should focus on improving quantitative readouts via paper substrate optimization and image-processing/deep learning methods, expanding validated genetic markers for fungicide resistance across pathogens, and enhancing microneedle robustness for diverse plant tissues (e.g., roots) to broaden sample compatibility.

• Current paper-based colorimetric readout is qualitative to semi-quantitative, with color nonuniformity from reagent diffusion limiting precise quantification; algorithmic color calibration and deep learning approaches are needed to improve quantitative performance.
• Resistance detection is constrained by the limited availability of validated mutation markers; only a subset of fungicide-resistant isolates can currently be identified, though coverage will improve as new markers are discovered.
• Microneedle-based extraction has so far been validated on leaves; tougher microneedles may be required for harder tissues (e.g., roots).
• While powerful, the assay’s performance relative to sequencing-based tools (e.g., MARPLE) involves trade-offs: the present method avoids amplification and costly instruments but does not provide comprehensive strain-level surveillance or broad genotyping without prior probe design.