Resazurin workshop on Candida albicans south of Marrakech. The bottleneck wasn't the assay, it was data entry. One app, thirty minutes, unblocked.
Last month I was invited to run a one-day workshop at EMINES, the engineering school in renewable energy and mineral engineering, south of Marrakech on the Benguérir campus. The professor who reached out, A., warned me by email: "Saliha, we have 12 M1 students, one spectro, and a shared Nikon for plates. Do what you can."
I accepted. Without fully grasping what "one spectro" meant.
Arriving at the lab
I arrived the night before, by train from Rabat then taxi from Marrakech. The campus is new, clean, ambitious. The biotech teaching lab where we would work is a sharper contrast: decent bench, working fume hoods, but the instrumentation is old. The spectrophotometer is a single-beam Shimadzu UV-1800, acquired in the late 1990s according to the inventory tag. It works. But it takes 2.5 seconds per reading and there is no microplate reader. So cuvette by cuvette, manually.
To document 96-well microplates from the resazurin assay, they use a Nikon D5000 mounted on a homemade stand built from a tripod screwed onto a wooden board. One single Nikon. For twelve students split into six pairs. And the Nikon USB cable plugs directly into the shared PC at the back, also unique. A. was clear-eyed: "Saliha, the bottleneck isn't the assay. It's data entry."
The workshop plan
Resazurin viability assay on Candida albicans ATCC 10231. Each pair tested a methanolic extract of a Moroccan aromatic plant at six concentrations in triplicate, plus positive (fluconazole 16 µg/mL) and negative (DMSO 1%) controls. Greiner 96-well black plates with clear bottom (A. had picked them up from a leftover order from the previous year).
The extracts had been prepared the week before by the students themselves: Salvia officinalis, Origanum compactum, Mentha pulegium, Rosmarinus officinalis, Thymus vulgaris, Lavandula angustifolia. Classic Moroccan repertoire, but with real diversity of origin (Atlas, Rif, Atlantic plain). These extracts have been tested thousands of times in the literature, except that for these students it was their first real fungal growth inhibition experiment, on their strain, in their lab.
The morning: the wall
We started at 8:30 am. Inoculation 30 min, addition of 0.01% resazurin, 4h incubation at 30°C in the oven. During those four hours, I ran a theory session on resazurin reduction to resorufin by viable cells and on best practices for reading color shifts (blue → pink), with the usual false-positive traps.
At 1 pm, plates out of the oven. They looked good, clean blue-to-pink gradients, you could already tell by eye that three extracts stood out. And then the wall. The ritual kicked in: pair 1 carries their plate to the Nikon, takes the photo, unplugs pair 2 from the PC, plugs in the USB cable, imports the photo, opens it in Windows photo viewer, reads the wells, dictates the values to their teammate who types them into Excel. Average time per pair: 22 minutes. Six pairs. Quick math: 2h12. Except that at 11 am they had not even started data entry. At 11 am, only three pairs had been through. End of session was at 4 pm.
A. looked at me. I know that look. It's the look of a teacher who knows her session will stop midway and that she will have to reschedule half a day she doesn't have on her calendar.
My phone, PhytoNote, and live capture
I had my Pixel 7a with me. And PhytoNote on it, already deployed in an M2 lab session in Rabat the month before (I told that story in the fifteen-plates story). I clipped my phone in as an additional camera above the plate, and the pairs entered their readings directly into PhytoNote while they observed the wells. Instead of transcribing the photo later from the PC, it's instantaneous: one student reads the colour of well A1, their teammate types the ordinal value (full blue, mauve, full pink, intermediate) into the app's 96-cell grid, and they move on to A2.
First try with pair 4 (pennyroyal mint extract). All 96 wells read by two: one reading, one typing. The pair ran a cross-check pass, corrected two borderline wells, validated. Local CSV export, Bluetooth transfer to the shared PC. Pair 4 total time: 6 minutes.
Pair 5: 4 minutes (they had picked it up). Pair 6: 4 minutes. Meanwhile pairs 1, 2, 3 finished their Nikon entry. By 1:30 pm everyone had their usable numerical results. A. said nothing. But she took a picture of my phone screen three times during the session, at moments when she thought I wasn't watching.
The biology
Of the six extracts, three showed significant inhibition of C. albicans viability at 250 µg/mL: Salvia officinalis (78% reduction of resorufin fluorescence), Origanum compactum (89%, the most active), and Mentha pulegium (71%). The other three stayed below the 30% threshold and were not carried forward. These results are consistent with what the literature reports on Lamiaceae essential oils, and Moroccan compact oregano in particular is known for its carvacrol and thymol content at levels that make it a serious candidate in antifungal studies.
But the pedagogical point isn't the science. It's that we finished the session with six pairs who each had their full dataset, their analysis, their dose-response curves sketched by hand on the board, and thirty minutes of discussion on possible biases. Instead of a session ending at 4 pm with four pairs stuck.
Coffee afterwards
We had coffee at the campus cafeteria. A. asked whether the app was publicly available and whether she could ask her department to bring it into the L3 biotechnology curriculum next year. I said yes, the stable version would be ready for the new term, and that I would come train her colleagues for a half-day if she wanted.
She said something that stayed with me: "You know Saliha, we have EMINES opening new renewable energy labs, but our biotech instrumentation, we inherited it from a French university equipment transfer in the early 2000s. We make do. When a data entry tool moves the timeline, that's not a gadget."
I took the train back to Rabat that evening. On board, I wrote a short note in PhytoNote: "EMINES workshop, 12 students, 6 pairs, photo mode validated under real conditions, L3 integration under discussion." It's the shortest and most satisfying note I've written this year.
The fast-capture mode and the 96-well grid architecture are described in the PhytoNote technical note, and the product page lists the supported assays. If you teach M1 or L3 students and your lab faces a similar bottleneck to the one at EMINES, the code is on GitHub and issues are open.