· 6 min read
What to look for in a field book app for plant breeding
A practical checklist for choosing a field book app: offline scoring, BrAPI and Field Book compatibility, trait ontologies, pedigree, photos and owning your data.
Paper field books and spreadsheets work until they don't. A sheet gets wet. A column gets pasted one row off. A score typed in December doesn't match what was written in July, and nobody can say who changed it.
A field book app should make scoring faster and the data cleaner. Plenty of apps manage one and not the other. This checklist covers what matters most for plant breeding trials, and how to test each point before you trust a season to it.
1. It works with no signal
Research stations are full of dead zones. A field book app that needs a connection will fail on the day you need it most.
Look for an app that opens with no signal, keeps the field books you've opened, saves scores on the device and syncs later by itself. It should also show clearly what hasn't been sent yet.
Test it: put a phone in airplane mode, score twenty plots, close the app, open it again, then turn the signal back on. Every score should arrive, once.
2. It's fast in the plot
Scoring speed is plots per hour, and small things add up.
- Plots in walking order, including serpentine layouts.
- Big targets for 1–9 scores, and a number pad for measurements.
- Jumping to a plot by scanning the barcode on its stake.
- Voice entry, so your eyes stay on the plot.
- Labels and barcodes printed from the same system, so they always match.
3. Traits are defined, not just named
"Lodging" means little without a scale. A good app stores each trait as a trait, a method and a scale: what you observe, how you observe it, and in what units or categories. That's the same structure the Crop Ontology uses.
The app should check every score against its trait: inside the range, a whole number when it should be, one of the allowed categories. A typo of 45 on a 1–9 scale should be caught in the field, not in the analysis.
Better still, the definition travels with the data: when to score (the growth stage), what each score looks like in words and photos, and which version of the protocol a score was taken under. If several people score, some way to check that they agree is worth a lot.
4. Trait ontologies
The Crop Ontology publishes standard trait definitions for many crops, each with a stable ID. Linking your traits to those IDs makes your data easier to share, to combine with other programs' data, and to read in ten years.
You don't need to give up your own trait names. You need the app to store the ontology ID beside each trait, and to export a trait dictionary.
5. BrAPI and Field Book compatibility
BrAPI, the Breeding API, is an open standard that lets breeding tools talk to each other. A system that speaks BrAPI can send trials, plots, lines and traits to a field app and take the scores back, with no files in between.
Field Book, the open-source Android app from PhenoApps, is widely used for scoring and can work this way. GridScore is another field app with BrAPI support. If your team already uses one of them, check:
- which BrAPI version the system supports, and whether it can take scores back, not just send trials out;
- how a phone signs in: a key pasted in, or a sign-in page;
- what happens to a score that fails a check: refused with a reason, or saved anyway;
- whether importing and exporting files still works, for the day the connection doesn't.
6. Pedigree and germplasm
Trials are only half the story. The app should know each line's pedigree, written in a standard notation like Purdy (A/B//C), with crosses, generations and seed lots behind it. Then every score sits on a line with a history, and you can ask how related two lines are before you cross them.
7. Photos
Photos settle arguments about scores and help train new scorers. Check that a photo is tied to the plot (and ideally the trait), what happens when you take one with no signal, and who can see it.
8. Checks and history
Mistakes happen. What matters is that they're caught, and that the record shows what changed. Look for flags on scores that don't fit, such as a value far from the line's other plots or a run of identical values from a slipped finger. Look for a score history that keeps who changed what, and when. Nothing should be overwritten silently.
9. Your data, and getting it out
Check this one before you sign up, not after.
- Can you export what you put in, in open formats like CSV or Excel?
- Is there an API (BrAPI counts) for getting data out with your own scripts?
- Who else can see your data? Can a technician enter scores without being able to change trial designs?
- What happens to the data if you stop using the app?
Test it: load a real trial, score it, and export it. Open the export somewhere else. If that's hard now, it'll be harder with ten years of data in it.
A short checklist
| Ask | Why it matters |
|---|---|
| Does it open and save scores with no signal? | Dead zones are normal at research stations |
| Does it check scores against each trait's scale? | Typos caught in the field, not in December |
| Can traits carry Crop Ontology IDs? | Data you can share and combine |
| Does it speak BrAPI, both ways? | Field Book and other tools without files |
| Does it know pedigrees? | Scores on lines with a history |
| Is every change kept with who and when? | A record you can trust |
| Can you get your data out? | Your data stays yours |
How Crossling fits
Crossling is a field book built around this list.
- Offline. It installs on a phone or tablet from the browser, like an app. It opens with no signal and keeps the field books, trait bible and task list you've opened. Scores and ticked-off tasks wait on the device until there's signal, and so do photos taken to measure a score.
- Fast scoring. One plot at a time in walking order (serpentine included), big buttons for 1–9 scores, voice entry, barcode scanning, photos of a plot while you score it, and a map mode that follows your GPS. Plot labels with barcodes print from the same trial.
- Defined traits. A library of ready-made traits for wheat, barley, canola, pulses, corn and other crops, or your own: measurements with units and ranges, 1–9 scores with what each score means, categories and dates. Every score is checked against its trait.
- One way to score. The trait bible holds each trait's meaning, timing and reference photos, in agreed, numbered versions. Calibration sessions show who scores high or low.
- Ontologies. A trait can carry its Crop Ontology ID, and the trait dictionary exports in the Crop Ontology layout.
- BrAPI v2.1. Field Book and GridScore can pull trials, plots, lines and traits, and send scores back. Sign in with Field Book's Authorize button, or with an access key from the Connect page. Scores sent this way go through the same checks as scoring in Crossling. One that fails isn't saved: Field Book keeps it as not sent, and the reason shows on the Connect page. Field Book's file export and import work too. The developer docs list every call, for your own scripts.
- Pedigrees. Every kind of cross, pedigrees written in Purdy notation, family trees, and the coefficient of parentage between any two lines.
- Checks and history. Scores that don't fit are flagged, and every correction is kept in the plot's score history.
- Your data. A program's trials, scores and photos are seen only by its members, with roles for admins, breeders, technicians and viewers. Field books, results and the trait dictionary download as Excel, and BrAPI gives your own tools access to trials and scores.
Try the demo to score a plot in a real-looking spring wheat program, or create an account and start with one trial.