Abstract
The characterization of antimicrobial susceptibility and other relevant phenotypes in large collections of microbial isolates is a common need across research and clinical microbiology laboratories. Robotization provides unprecedented throughput but involves costs that are prohibitive for the average laboratory. Here, using affordable materials and open-source software, we developed Q-PHAST (Quantitative PHenotyping and Antimicrobial Susceptibility Testing), a unique solution for cost-effective, large-scale phenotyping in a standard microbiology laboratory. Single colonies are grown in a deep 96-well master plate, from which diluted aliquots are used to generate 96 spots on different experimental plates containing solid medium with the substance and concentration of interest. These plates are incubated on inexpensive flatbed scanners that monitor the growth of each spot by obtaining images every 15 min. A simple, python-based software, which can be used via a graphical interface on various operating systems (https://github.com/Gabaldonlab/Q-PHAST), analyzes the images to infer growth, fitness (e.g., doubling rate) and susceptibility (e.g., minimum inhibitory concentration) measures. With <120 min of hands-on time per day for three consecutive days, ready-to-use results are obtained and presented in tables or graphs. This solution enables non-experts with limited resources to perform accurate quantitative phenotyping on hundreds of strains in parallel.
Key points
-
Q-PHAST enables quantitative and temporal analysis of microbial phenotypes, including antimicrobial susceptibility testing, for hundreds of strains in parallel by using scanners to monitor the growth of spots on solid media and software to analyze them through a graphical interface.
-
It is a unique solution for cost-effective, large-scale phenotyping in a standard microbiology laboratory using affordable materials and open-source software.
This is a preview of subscription content, access via your institution
Access options
Access Nature and 54 other Nature Portfolio journals
Get Nature+, our best-value online-access subscription
$32.99 / 30 days
cancel any time
Subscribe to this journal
Receive 12 print issues and online access
$259.00 per year
only $21.58 per issue
Buy this article
- Purchase on SpringerLink
- Instant access to full article PDF
Prices may be subject to local taxes which are calculated during checkout
Similar content being viewed by others
Data availability
All growth-per-time point measurements related to the analyses and figures shown here are found in the folder ‘misc/Nunez-Rodriguez_2024_growth_measurements’ of https://github.com/Gabaldonlab/Q-PHAST. The fitness, relative fitness and susceptibility measurements related to the strains used here are in Supplementary Tables 3–5. In addition, the folder ‘testing/testing_subsets’ within our GitHub repository (https://github.com/Gabaldonlab/Q-PHAST) contains examples of the images and PLs that can be analyzed with Q-PHAST.
Code availability
Q-PHAST’s code and documentation can be found in https://github.com/Gabaldonlab/Q-PHAST. We used the release v1, available at https://github.com/Gabaldonlab/Q-PHAST/releases/tag/v1, to generate all the results and figures in this protocol.
References
Kamrad, S. et al. Pyphe, a python toolbox for assessing microbial growth and cell viability in high-throughput colony screens. eLife. 9, e55160 (2020).
Kritikos, G. et al. A tool named Iris for versatile high-throughput phenotyping in microorganisms. Nat. Microbiol. 2, 17014 (2017).
Barton, D. B. H. et al. PHENOS: a high-throughput and flexible tool for microorganism growth phenotyping on solid media. BMC Microbiol. 18, 9 (2018).
Wagih, O. et al. SGAtools: one-stop analysis and visualization of array-based genetic interaction screens. Nucleic Acids Res. 41, 591–596 (2013).
Holland, I. & Davies, J. A. Automation in the life science research laboratory. Front. Bioeng. Biotechnol. 8, 571777 (2020).
del Olmo, V. et al. Origin of fungal hybrids with pathogenic potential from warm seawater environments. Nat. Commun. 14, 6919 (2023).
Mixão, V. et al. Evolution of loss of heterozygosity patterns in hybrid genomes of Candida yeast pathogens. BMC Biol. 21, 105 (2023).
Ksiezopolska, E., Schikora-Tamarit, M. À., Nunez-Rodriguez, J. C. & Gabaldón, T. Long-term stability of acquired drug resistance and resistance associated mutations in the fungal pathogen Nakaseomyces glabratus (Candida glabrata). Front. Cell. Infect. Microbiol. 14, 1416509 (2024).
Addinall, S. G. et al. Quantitative fitness analysis shows that NMD proteins and many other protein complexes suppress or enhance distinct telomere cap defects. PLoS Genet. 7, e1001362 (2011).
Klemm, C., Howell, R. S. M. & Thorpe, P. H. ScreenGarden: a shinyR application for fast and easy analysis of plate-based high-throughput screens. BMC Bioinformatics 23, 60 (2022).
Zackrisson, M. et al. Scan-o-matic: high-resolution microbial phenomics at a massive scale. G3 6, 3003–3014 (2016).
Wagih, O. & Parts, L. gitter: a robust and accurate method for quantification of colony sizes from plate images. G3 4, 547–552 (2014).
Doherty, H. M., Kritikos, G., Galardini, M., Banzhaf, M. & Moradigaravand, D. ChemGAPP: a tool for chemical genomics analysis and phenotypic profiling. Bioinformatics 39, btad171 (2023).
Bischof, L. et al. Spotsizer: high-throughput quantitative analysis of microbial growth. Biotechniques 61, 191–201 (2016).
Levin-Reisman, I., Fridman, O. & Balaban, N. Q. ScanLag: high-throughput quantification of colony growth and lag time. J. Vis. Exp 2014, 51456 (2014).
Tegally, H., San, J. E., Giandhari, J. & de Oliveira, T. Unlocking the efficiency of genomics laboratories with robotic liquid-handling. BMC Genomics 21, 729 (2020).
Lawless, C., Wilkinson, D. J., Young, A., Addinall, S. G. & Lydall, D. A. Colonyzer: automated quantification of micro-organism growth characteristics on solid agar. BMC Bioinformatics 11, 287 (2010).
Banks, A. P., Lawless, C. & Lydall, D. A. A quantitative fitness analysis workflow. J. Vis. Exp. 2012, 4018 (2012).
Narayanan, S. et al. Quantitative fitness analysis identifies exo1Δ and other suppressors or enhancers of telomere defects in Schizosaccharomyces pombe. PLoS ONE 10, e0132240 (2015).
Berman, J. & Krysan, D. J. Drug resistance and tolerance in fungi. Nat. Rev. Microbiol. 18, 319–331 (2020).
Ksiezopolska, E. et al. Narrow mutational signatures drive acquisition of multidrug resistance in the fungal pathogen Candida glabrata. Curr. Biol. 31, 5314–5326 (2021).
Daniel, B. & Lawless, C. Modeling competition for nutrients between microbial populations growing on a solid agar surface. Preprint at bioRxiv https://www.biorxiv.org/content/10.1101/086835v1 (2017).
Acknowledgements
The T.G. group acknowledges support from the Spanish Ministry of Science and Innovation for grants PID2021-126067NB-I00, CPP2021-008552, PCI2022-135066-2 and PDC2022-133266-I00, cofounded by ERDF ‘A way of making Europe’; from the Catalan Research Agency (AGAUR) SGR01551; from the European Union’s Horizon 2020 research and innovation programme (ERC-2016-724173); from the Gordon and Betty Moore Foundation (grant GBMF9742); from the ‘La Caixa’ Foundation (grant LCF/PR/HR21/00737); and from the Instituto de Salud Carlos III (IMPACT grant IMP/00019 and CIBERINFEC CB21/13/00061- ISCIII-SGEFI/ERDF). J.C.N.-R. received a predoctoral fellowship from the Spanish Ministry of Science and Innovation (grant number PRE2019-088193). M.A.S.-T. received a predoctoral fellowship from the ‘La Caixa’ Foundation (grant LCF/BQ/DR19/11740023). Some of the figures were created with BioRender.com The authors thank all of the members of the T.G. group for key support during this work. The authors particularly thank V. d. Olmo, I. Ahaik, D. Fuentes, E. Saus and M. Bernabeu, who provided useful feedback, which was key for the project development.
Author information
Authors and Affiliations
Contributions
J.C.N.-R. contributed to the conceptualization of the experimental design and software; developed all parts of the wet laboratory protocol; performed all experiments, analysis and validation; and contributed to writing the protocol. M.A.S.-T. contributed to the conceptualization of the experimental design and software, developed the entire pipeline and graphical interface, wrote the GitHub documentation and contributed to writing the protocol. E.K. contributed to the conceptualization of the experimental design and participated in the experimental validation. T.G. supervised the entire process, acquired the funds, contributed to the conceptualization of the experimental design and software and contributed to writing the protocol. All the authors read and approved the final manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Reviewer recognition
Nature Protocols thanks the anonymous reviewers for their contribution to the peer review of this work.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Key references
del Olmo, V. et al. Nat. Commun. 14, 6919 (2023): https://doi.org/10.1038/s41467-023-42679-4
Ksiezopolska, E. et al. Front. Cell. Infect. Microbiol. 14, 1416509 (2024): https://doi.org/10.3389/fcimb.2024.1416509
Mixão, V. et al. BMC Biol. 21, 105 (2023): https://doi.org/10.1186/s12915-023-01608-z
Extended data
Extended Data Fig. 1 Plate layouts (PLs) used in this protocol.
a, Distribution of strains in the Master Plate (MP) design. The MP is divided into 4 blocks, with each block containing one biological replicate of each of 20 strains, plus 4 controls distributed in a predetermined order. In the first block, starting from well A1, the strains are arranged in order until the last strain in the block is reached at D6. Following the same order, the replicates of the fourth block are placed starting at E7 and ending at H12. In the remaining two quadrants, the second and third, the strains are placed in reverse order, creating a mirror distribution to their parallel quadrant. This logical distribution across different positions on the plate helps to avoid potential position bias. This distribution is automatically generated when the list of strains to be tested is entered in the PL (Supplementary Table 6). Other distributions are possible by using the PL_EmptyFree_To_Use.xlsx template (Supplementary Table 7). b, Distribution of strains in all experiments presented in this protocol following the previously proposed layout. c–e, Configuration of the PLs used in this protocol, where c shows the fitness experiment, d shows the stress response experiment and e shows the antifungal susceptibility test (AST) against anidulafungin. The first panel indicates the name of the experiment. The second reflects the scanners used and the composition of each plate for each scanner. The third indicates the compound concentrations on each plate; if a value is 0, it is used as a growth control reference from which relative values are calculated. Note that you can find the PL files in https://github.com/Gabaldonlab/QPHAST/blob/main/misc/plate_layouts_Q-PHAST_study.xlsx.
Extended Data Fig. 2 Q-PHAST graphical user interface.
a, Pipeline initialization command and initial graphical window for operating system (OS) selection. b, Window to choose the docker image to be used; at launch only version 1 is available. This window will allow users to choose among the installed Q-PHAST docker images. c, Folder selection window where the ‘input’ folder is selected and where you want the ‘output’ folder with all the results to be created. Click the ‘Run’ button to proceed to the next screen. d, Optional parameter settings. Adjust the experiment duration, commonly set at 24 h but adaptable on the basis of experimental requirements. Experimental images provided should have been acquired at this duration or longer. In addition, you can set the minimum nAUC value (raw fitness estimate) that is required for a spot to be considered growing, crucial for susceptibility analysis. Lastly, the ‘enhance contrast’ configuration improves image quality for optimal software performance. However, this may also be deactivated by setting ‘enhance contrast’ to ‘False’, which could be useful if the analyzed images have sufficient contrast and the default parameters yield poor results (this may be assessed by empirically looking at the pipeline’s outputs). e, Option for verification of spot coordinates and bad spots. Choose ‘Yes’ for reliable results, where the user is asked to verify the coordinates of all plates and the bad spots. The ‘No’ option performs these steps automatically and is recommended only for testing and development purposes. f, Definition of spot coordinates. Select the first spot (top left) and the last (bottom right) by clicking with the mouse and pressing ‘Enter’. To repeat the selection, double-click. Accurate identification of these spots is necessary for subsequent definition of the coordinates. g, Verification of spot coordinates. Press ‘Y’ (Yes) to accept verification and ‘N’ (No) to reject. If rejected, you will return to the step of defining coordinates (see f). h, Verification of predicted bad spots. The software identifies as ‘bad spots’ those that are outliers within the fitness distribution of different replicates of a given strain. For each strain in a given plate plate, we considered Q1 (first quartile), Q3 (third quartile) and IQR (interquartile range, Q3 − Q1) of the nAUC values across technical replicates. Potential ‘bad spots’ are defined as those that are outside of the range (Q1 − 2.5 × IQR, Q3 + 2.5 × IQR). An informative window displays three images summarizing spot growth on the plate, highlighting potential bad spots in a red box and their replicates in black. Growth curves are also shown in red for the potential bad spot and in black for its replicates. Identify the potential bad spot as bad by pressing ‘b’ or as good by pressing ‘g’. Bad spots will be excluded from future calculations and reflected in ‘extended_outputs’ in ‘bad_spot.xlsx’. I, Terminal showing the log of the pipeline. At the beginning, the complete execution code will be presented for reproducibility. This information will also be available in the folder ‘extended_outputs’. At any point in the process, you can see which step the pipeline is in, and upon completion, you will find a ‘success’ message and the time taken to complete the process.
Extended Data Fig. 3 Complete outputs generated by Q-PHAST.
a, Tree folder structure of the outputs generated by QPHAST. Results are organized into two main blocks: one containing simple result tables and summary plots, and the other containing extended outputs. Within extended outputs, there are folders with various graphs generated by using different fitness estimates. It also includes comprehensive, per-spot results with all growth, fitness and susceptibility calculations. In addition, the plate_layout used, the list of bad spots (both manually identified and automatically detected), the end report and a reduced_input folder containing information necessary for replication and troubleshooting are present. b–f, Examples of graphs generated in the outputs. b, Heatmap displaying the fitness of each strain under different drug concentrations. The circles inside each heatmap cell indicate the median absolute deviation among replicas to visualize dispersion. In this example plot, the median nAUC is the used fitness estimate, but Q-PHAST outputs equivalent heatmaps for other fitness estimates (Supplementary Methods). c, Line graph showing the fitness (nAUC for this example) of different replicates of a strain under each drug concentration. d, Heatmap of strain susceptibility measured in rAUC, MIC50 and SMG. This visualization is equivalent to Fig. 1g. e, Example of a graph found within the ‘Growth_curves_and_images’ folder, illustrating growth curves of each replica over time at different concentrations. Three images at the bottom of each graph show growth at different times, with each replica highlighted in a square of the corresponding color shown in the upper graph. This is useful for quality control. f, Growth curves for each spot on the plates, indicating the strain, modeling its growth and displaying various fitness estimates. This visualization is equivalent to Fig. 2b.
Extended Data Fig. 4 Technical and biological reproducibility of QPHAST.
Boxplots showing, for each strain, the distribution across biological replicates of fitness (nAUC) in different YPD plates (a), relative fitness (nAUC_rel) when grown on various conditions (b) and anidulafungin susceptibility (rAUC) by different fitness estimates (c). Boxes show the IQR of the distribution, from Q1 to Q3, while the line represents the median. The whiskers extend to points that lie within 1.5 × IQR of the first and third quartiles, and values outside this range are shown as independent points.
Extended Data Fig. 5 Growth across time for all replicates grown in YPD from the ‘assessment of stress susceptibility’ experiment, for samples Cg-wt1 (a) and Ct-wt (b).
The left panel shows the spot of each replicate across 2-h time points from processed images, which are analyzed within Q-PHAST to obtain cell density (growth) measurements (shown in the right panel). We show these two strains to showcase how Q-PHAST successfully addresses two potential biases of spot image analysis. First, there are some replicates in Cg-wt1 (a) grown close to the plate border (A5, A8), which does not affect growth measurements. Second, the morphology of the colonies in Ct-wt (b) is slightly less rounded, but this does not have a significant impact on growth inferences.
Extended Data Fig. 6 Growth of different species and their replicates when following the Q-PHAST methodology.
a–c, Different yeast species: Nakaseomyces bacilisporus (a), Pichia cactophila (b) and Saccharomyces uvarum (c). d–f, Different bacterial species: Escherichia coli DH5 alpha (d), a clinical strain of Klebsiella CMA2 (e) and a clinical strain of Stenotrophomonas SNB_C1 (f), were grown from single colonies in an MP. The yeast MP contained YPD, whereas the bacterial MP contained LB. According to the protocol, they were inoculated into the EP with YPD agar or LB agar, respectively, and incubated on the scanners for 48 h for yeast and 24 h for bacteria, both at 30 °C. The image shows the automated results generated by Q-PHAST, provided in extended_outputs/growth_curves_and_images, where the growth of these strains at three time points can be observed along with their growth curves. Note that there are a few spots incorrectly boxed (e.g., the red square (D9) in N. bacilisporus (a) or the orange square (B12) in P. cactophila (b), where growth is only inferred within the box. This may lead to slightly inaccurate growth measurements, which we consider as an unavoidable limitation of such image-based growth measurements. Although this is an infrequent artefact, it should be taken into account. We encourage users to check out such plots for their experiments, to be able to flag as ‘bad spots’ (in the input PL) the incorrectly placed ones.
Extended Data Fig. 7 Scanner setup instructions.
This figure illustrates the assembly steps for establishing the system outlined in the protocol. a, Unpacked Epson V600 photo scanner. b, The scanner with the lid removed and positioned at the bottom, leaving the scanner surface exposed. For ease of handling, it can be secured with black adhesive tape. c, A template indicating the positions of the plates on the scanner is presented. With the front buttons on the left, the first plate corresponds to the upper left, the second to the upper right, the third to the lower left and the fourth to the lower right. d, By using the template from c, an EVA foam template is cut to the dimensions of the omnitray plates. The cut sections serve as opaque covers for each plate. e, The template from the previous step is placed on the scanning surface to enhance plate placement, keeping them fixed and immobile. The non-plate area remains entirely black, improving image capture. f, Three scanners, prepared as described previously, are arranged inside a temperature-controlled incubator, with cable connections accessible externally. g, Final assembly of three scanners within a temperature-controlled incubator. At the top is the computer that controls and stores images from the scanners. The scanners connect to this computer via USB. Both cables and scanners are labeled with colors for traceability. Scanner power cables are organized on the opposite side for orderly connection to the power supply. The surfaces of the incubator exposed to light are covered to ensure total darkness when the incubator is closed, thereby improving image quality and homogeneity.
Supplementary information
Supplementary Information
Supplementary Methods and Supplementary Tables 2, 3, 4 and 5
Supplementary Table 1
Comparison table between different large-scale phenotyping methods and Q-PHAST and its advantages and disadvantages
Supplementary Table 6
Empty PL document that is necessary to run Q-PHAST. This document orders the MP with 24 strains as proposed in the protocol
Supplementary Table 7
Empty PL required to run Q-PHAST, but unlike Supplementary Table 6, this PL allows the user to arrange the MP as desired
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Nunez-Rodriguez, J.C., Schikora-Tamarit, M.À., Ksiezopolska, E. et al. Simple large-scale quantitative phenotyping and antimicrobial susceptibility testing with Q-PHAST. Nat Protoc (2025). https://doi.org/10.1038/s41596-025-01179-z
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41596-025-01179-z
