Source
doi:10.3390/foods12193636 — original publication (opens in a new tab; the file is not redistributed)
Summary
Peer-reviewed MDPI Foods study (Kabutey, Herák & Mizera, 2023, 12(19):3636, open access CC-BY, DOI 10.3390/foods12193636) from the Czech University of Life Sciences Prague that runs 200 g batches of six oilseeds — sesame, pumpkin, flax, milk thistle, hemp and cumin — through a small tabletop electric cold screw press (motor 180 W, oil temperature ≤40 °C) and measures oil yield, seedcake, sediments and losses. Its single load-bearing contribution for Neobiome Intelligence is the oil expression efficiency: defined explicitly as oil yield ÷ oil content of the seed, this is exactly the oil-in-seed → oil-in-bottle recovery fraction the food model needs. Measured on this cold press the efficiencies were sesame 48.57–82.70%, flax 63.02–72.97%, hemp 53.45–81.24%, milk thistle 48.2–72.3%, pumpkin 20.89–60.77% (cumin excluded — too little oil, 1.16–2.22% content). It is a transferable proxy (press physics, not location): the number describes what a small screw press does to a seed, so it carries across to NZ crops even though the trials were run in Czechia. The finding the model must act on is that counting oil-in-seed as usable oil overstates extractable oil by roughly 15–50% on a community-scale cold press; hemp — a Neobiome target crop — is directly covered at ~53–81% recovery.
Neobiome Intelligence relevance
- Resolves RT_391 (screw-press oil recovery fraction). Supplies the missing multiplier in the food model’s oilseed → oil pathway: usable oil = (seed yield) × (oil-in-seed fraction) × (press recovery fraction). Before this, the model counted the second term as if it were the third (i.e. treated oil-in-seed as bottled oil), overstating oil output by ~15–50% LIT_122.
- Model cell it feeds — the oilseed press-recovery discount in D02. Recommended working band on a small cold screw press ~50–85% (mid ~70%) for the general oilseed pathway; hemp-specific ~53–81% is directly applicable because hemp is a target crop. The band is a synthesis of the verbatim ranges (sesame/flax/hemp cluster 48–83%), flagged as a recommendation, not a source constant.
- Multiplies the existing oil-in-seed sources (cross-ref). The recovery fraction is the term that was missing on top of the corpus’s oil-in-seed figures: hemp seed 950–1,800 kg/ha at 25–38% oil OT_168; oilseed rape 2.4 t/ha at 42% oil ≈ ~1,008 kg oil in the seed LIT_095, cold-pressed in NZ with no solvent/no heat OT_183. Those pages give oil-in-seed; this page gives the fraction of it a cold press actually recovers. The OSR ~1,008 kg/ha figure in particular must now be de-rated by ~50–85% before it counts as usable oil LIT_122.
- ⚠ Proxy scope + scale caveat. Czech trials, six species (only hemp and, loosely, OSR-adjacent oilseeds map to NZ crops). The press is a small tabletop unit — the authors state a community would need medium/large presses to push recovery toward the top of the band; industry hits <3% residual oil only with pre-pressing + solvent extraction, which an off-grid community will not run. So for a small cold press the conservative floor (~50%) is the right lower bound for not overstating oil — the measured minima on this same press were 48–53% (milk thistle 48.2, sesame 48.57, hemp 53.45); the ~85% top is only reachable with a bigger press LIT_122.
Research targets
Documents to retrieve
- RT_391 — RESOLVED → LIT_122. Screw-press oil recovery fraction (oil-out ÷ oil-in-seed) for the food model’s oilseed pathway. Closed by this source: hemp 53.45–81.24%, band ~50–85% (mid ~70%), conservative small-press floor ~50%, transferable proxy. No new RT opened (minimise-new-RTs) — the NZ-species calibration of the band is noted inline as a caveat, not a new target.
Notes
- Capture format: the raw is the Europe PMC full-text XML for this article (PMC10573014), the open-access full text, rather than the MDPI HTML/PDF. It is an authoritative direct capture (not an AI synthesis), so no
## Retrieval provenanceblock applies. Section titles (not page numbers) are the citable anchors in the XML —source_locationcites the article’s numbered sections. - Efficiency denominators are literature values, not this study’s measurements. The oil-in-seed content ranges (hemp 25–38%, flax 38–44%, etc.) are drawn from the papers Kabutey et al. cite in §2, then divided into their own measured oil yields to produce the efficiency ranges. That is why each efficiency is a range, not a single number — it reflects the spread of literature oil-in-seed content.
- Cumin excluded from the efficiency calc (oil content only 1.16–2.22%, yield 3.46%) — not relevant to Neobiome.
- The core model message: oil-in-seed overstates usable oil. A cold press on this hardware recovered ~48–83% of the seed’s oil, so the food model must apply the recovery fraction (~50–85%, hemp ~53–81%, conservative floor ~50%) rather than assuming full extraction.
Connections
Links to
Sources (3): LIT_095 · OT_168 · OT_183
EDT domains (1): D02: Smart Food Systems & Agriculture
Referenced by
EDT domains (1): D02: Smart Food Systems & Agriculture