Research programme
The scientific case for gas-signal shelf-life forecasting
This page sets out the hypothesis, methodology, produce selection rationale, and current programme status for institutional funders and research partners considering engagement with SciFruVeg Technologies.
The hypothesis
Gas exchange as a leading indicator of senescence
Fresh produce undergoes predictable biochemical changes during senescence — the process of ripening and decay. These changes are accompanied by measurable shifts in the gas-exchange signature of the produce: rising CO₂ output as cellular respiration accelerates, and the production of ethylene and ethylene-related volatile organic compounds (VOCs) as the ripening cascade progresses.
The SciFruVeg hypothesis is that these gas-exchange signatures carry sufficient predictive signal to forecast remaining shelf-life with operationally useful accuracy — before visible decay, and without physical inspection of individual units.
If the hypothesis holds, a low-cost sensor array placed inside a storage chamber could provide a continuous, non-destructive estimate of remaining shelf-life for the produce held within it. This would allow operators to make evidence-based decisions about dispatch, redistribution, and waste reduction.
Produce selection
Four produce types. Four distinct senescence profiles.
The four produce types were selected for their commercial significance, their distinct gas-exchange and senescence characteristics, and the availability of food-science literature sufficient to support model development.
Apples
Well-characterised ethylene-driven ripening with an extensive post-harvest literature base. The predictable ethylene burst at the onset of ripening makes apples a strong candidate for model training and threshold calibration. High commercial volumes in both US and UK cold-chain operations provide access to operationally representative storage conditions.
Avocados
Rapid and commercially costly senescence window driven by a pronounced climacteric ethylene burst. The narrow window between optimal ripeness and decay makes accurate shelf-life forecasting particularly valuable. High loss rates in the supply chain create strong commercial motivation for validation partners.
Berries
Non-climacteric produce with CO₂ accumulation as the primary senescence signal. Berries test the model's ability to operate without an ethylene trigger, broadening the programme's applicability. High perishability and significant food-waste rates make them a commercially important test case.
Bananas
Highly predictable and well-documented ethylene burst at the onset of ripening. The regularity of the banana senescence profile makes it well-suited to initial model training and provides a controlled baseline against which the model's performance on less predictable produce types can be assessed.
Scientific grounding
Grounded in post-harvest biology
The programme draws on an established body of food-science literature on gas-exchange dynamics, climacteric and non-climacteric ripening, and the relationship between ethylene production and senescence rate. The four produce types under investigation were selected in part because the relevant literature is sufficiently developed to support model construction and validation design.
Ethylene biosynthesis and its role in climacteric ripening is well-characterised for apples and bananas, providing a strong baseline for model training and threshold calibration.
CO₂ accumulation as a primary senescence signal in non-climacteric produce (including berries) is documented in post-harvest physiology literature and informs the sensor specification for those produce types.
The rapid and commercially significant senescence window of avocados — driven by a climacteric ethylene burst — makes them a high-value test case for the forecasting model.
Methodology
Sense. Model. Validate.
The programme follows a three-stage methodology designed to move from sensor data collection through model development to institutional validation of forecast accuracy.
Gas sensing in storage
Low-cost CO₂ and ethylene-related VOC sensors are deployed inside cold-storage chambers at partner sites. Sensors monitor the gas-exchange signature of ripening produce continuously, capturing the biochemical progression of senescence as it unfolds in operational storage conditions. Sensor specifications are calibrated to the gas-exchange profiles of each produce type.
Shelf-life modelling
Sensor data feeds a machine-learning model grounded in post-harvest food-science literature. The model is trained to identify the gas-exchange patterns associated with different stages of senescence and to forecast remaining shelf-life from the current gas signature. Model architecture and training methodology are co-designed with research partners.
Institutional validation
Forecast accuracy is assessed against ground-truth shelf-life observations at partner sites. Validation partners provide the operational context, storage conditions, produce data, and scientific expertise needed to test whether the model generalises beyond controlled conditions. Validation criteria are agreed with research partners before data collection begins.
Programme status
Where we are now
Validate
Current stageSeeking research partners (post-harvest and food-science groups) and site partners (US and UK operators) to co-design the validation protocol, host sensor installations, and provide ground-truth shelf-life records.
Pilot
Next stageCommercial pilots in the United States and the United Kingdom, following completion of the validation work. Customers see evidence before they commit. A commercial pilot prospect is already in progress through Konero YellowBox.
Deploy
Future stageRepeatable rollout across additional produce types, geographies, and storage environments, informed by the findings of the validation and pilot phases.
Partnership
What we are looking for — and what partners receive
Research partners
Post-harvest and food-science groups to co-design the protocol, review the model, and co-author results.
What partners receive
- A say in experimental design and the validation criteria
- Access to results, and co-authorship where contributions warrant it
- Clear data terms, agreed in writing before work begins
Site partners
Operators and facilities able to host sensors and provide ground-truth shelf-life records.
What partners receive
- Early access to the system and a role in shaping the dashboard and sensor workflow
- Access to results relevant to their storage operations
- Clear data terms, agreed in writing before work begins
Funders
Innovation and research funders interested in food-waste reduction.
What partners receive
- A structured research programme with defined validation criteria and a clear commercial pathway
- Regular reporting against programme milestones
- Co-branding and acknowledgement in published outputs
Data ownership: Data ownership terms are discussed and agreed individually with each partner before any data collection begins. No data is collected or shared without a signed agreement.
Commercial pathway
From research to deployment via Konero YellowBox
SciFruVeg Technologies is developed under the Konero Group. The commercial pilot and deployment pathway runs through Konero YellowBox — Konero's applied innovation and commercialisation vehicle. This structure separates the research programme from commercial risk during the validation phase: research partners engage with the science, not the commercial outcome. The commercial pilots in the United States and the United Kingdom follow the validation work, so customers see evidence before they commit.
Interested in the programme?
We welcome enquiries from post-harvest research groups, Food Innovation Centres, cold-chain operators, and research funders. No commitment is required at this stage.
research@scifruvegai.com