{ "id": "co-fermentation", "url": "https://coffearium.org/entries/co-fermentation/", "license": "CC-BY-4.0", "title": "Starter cultures and co-fermentation", "summary": "Steering coffee fermentation by adding selected yeasts or bacteria, or by fermenting the coffee with fruit, spices or other ingredients – how strains are chosen, what trials found, and what competition rules allow.", "domain": "post-harvest-processing", "kind": "Method", "stage": "process", "topics": [ "science" ], "topicSections": {}, "lang": "en", "status": "verified", "reviewer": "Coffearium verification pass", "lastVerified": "2026-10-21T00:00:00.000Z", "updated": "2026-10-21T00:00:00.000Z", "added": "2026-10-21T00:00:00.000Z", "brief": [ "This entry uses “co-fermentation” for two practices – inoculating coffee with selected yeasts or lactic acid bacteria, or fermenting it together with fruit, spices or other ingredients.7,13,16", "Candidate starter strains are screened for pectin-degrading enzymes, aroma production and the absence of unwanted metabolites; some yeasts also inhibit ochratoxin A-producing moulds in the laboratory.2,3,4", "Inoculated trials change acids and volatile compounds and often score well, but score gains are not always statistically significant or separable from other changes.7,12,14", "Aroma compounds dissolved in fermentation liquid can enter depulped seeds, but the parchment slows some of them and part is metabolised.17,18", "The 2026 World Brewers Cup rules ban additives only after coffee reaches the green coffee stage; the 2023 Brazil Cup of Excellence Experimentals category allowed no additives other than water.21,22" ], "quickFacts": [ { "label": "Meaning 1", "value": "Inoculation with selected yeasts or lactic acid bacteria (starter cultures)" }, { "label": "Meaning 2", "value": "Fermentation with added fruit, spices, juices or other ingredients (“infused”)" }, { "label": "Yeasts tested", "value": "Saccharomyces cerevisiae, Torulaspora delbrueckii, Candida parapsilosis, Pichia spp., Hanseniaspora uvarum, Meyerozyma caribbica" }, { "label": "Inoculum in trials", "value": "10⁷ cells/g of coffee; 10⁷–10⁸ cells/mL" }, { "label": "Selection criteria", "value": "Pectinolytic enzymes, aroma, no unwanted metabolites, persistence, antagonism to moulds" }, { "label": "Competition boundary", "value": "No additives after the green coffee stage (World Brewers Cup 2026)" } ], "aka": [ "Co-fermentation", "Co-fermented coffee", "Infused coffee", "Inoculated fermentation", "Additive fermentation" ], "linkTerms": [ "co-fermented", "starter cultures", "yeast starter", "yeast starters", "inoculated coffee", "inoculation" ], "related": [ "fermentation", "microorganisms", "anaerobic-fermentation", "carbonic-maceration", "natural-process", "pulped-natural", "washed-process", "mucilage", "ochratoxin-a", "seed-physiology" ], "contested": [ "Effect of starter cultures on the cup score: some trials report higher scores for inoculated coffee5,15,23, while in a wet-fermentation trial the differences in sweetness, balance, clean cup, uniformity, aftertaste and overall score were reported as not statistically significant12, and in another trial the uninoculated control also scored above 85 points.14", "Mixed cultures of yeast and lactic acid bacteria: with ripe cherries, a combined inoculum raised sugar use and aroma complexity, but with immature cherries the combined treatment had no significant effect and only the yeast alone helped.13", "Rules for additives in competitions: the 2026 World Brewers Cup rules ban additives only from the green coffee stage onwards21, while the 2023 Cup of Excellence Experimentals category in Brazil accepted no additives other than water.22" ], "corrections": [ { "date": "2026-10-21T00:00:00.000Z", "change": "Removed the 53% inhibition figure for the 2017 yeast–mould study; the abstract and the results section of the paper give different values for it. The claim that Rhodotorula mucilaginosa inhibited ochratoxin A production by all isolates is now attributed to the abstract, with the contrary results noted.", "reason": "Internal inconsistency in source 4." }, { "date": "2026-10-21T00:00:00.000Z", "change": "Removed the statement that the Colombian preprint found significant differences only for uniformity and cleanness; a preprint is no longer used for any result.", "reason": "Source 16 is a non-peer-reviewed preprint; its peer-reviewed version (source 15) is used for results." }, { "date": "2026-10-21T00:00:00.000Z", "change": "The on-farm wet trial with Pichia fermentans used inoculation with or without 2% (w/v) sucrose, not always with sucrose; the sucrose detail is now cited to the original study.", "reason": "Abstract of source 11." }, { "date": "2026-10-21T00:00:00.000Z", "change": "World Brewers Cup rules updated from the 2024 to the 2026 edition (wording of the additive rule unchanged); the Cup of Excellence rule is no longer generalised beyond the 2023 Brazil programme.", "reason": "Currency of source 21; scope of source 22." } ], "sources": [ { "id": "s1", "title": "Coffee fermentation: Expedition from traditional to controlled process and perspectives for industrialization", "authors": "Elhalis, H.; Cox, J.; Zhao, J.", "year": 2023, "publisher": "Applied Food Research 3(1), 100253", "url": "https://www.sciencedirect.com/science/article/pii/S2772502222002086", "doi": "10.1016/j.afres.2022.100253", "tier": "A", "quote": "A wide range of endogenous isolates was detected with pectinolytic activities including bacteria such as Klebsiella, Erwinia, Bacillus and Lactobacillus and yeasts such as Pichia, Saccharomyces, and Candida.", "accessed": "2026-10-21T00:00:00.000Z" }, { "id": "s2", "title": "Coffee and Yeasts: From Flavor to Biotechnology", "authors": "Ruta, L. L.; Farcasanu, I. C.", "year": 2021, "publisher": "Fermentation 7(1), 9", "url": "https://mdpi.com/2311-5637/7/1/9", "doi": "10.3390/fermentation7010009", "tier": "A", "quote": "The selection of microorganism for the fermentation step is based on many criteria, such as PL secretion", "accessed": "2026-10-21T00:00:00.000Z" }, { "id": "s3", "title": "Isolation, Identification, and Characterization of Pectinolytic Yeasts for Starter Culture in Coffee Fermentation", "authors": "Haile, M.; Kang, W. H.", "year": 2019, "publisher": "Microorganisms 7(10), 401", "url": "https://www.mdpi.com/2076-2607/7/10/401", "doi": "10.3390/microorganisms7100401", "tier": "A", "quote": "The highest PDI was obtained from S. fibuligera (strain KNU18Y4) at 178%.", "accessed": "2026-10-21T00:00:00.000Z" }, { "id": "s4", "title": "Use of wild yeasts as a biocontrol agent against toxigenic fungi and OTA production", "authors": "de Souza, M. L.; Passamani, F. R. F.; Ávila, C. L. da S.; Batista, L. R.; Schwan, R. F.; Silva, C. F.", "year": 2017, "publisher": "Acta Scientiarum. Agronomy 39(3), 349–358", "url": "https://www.redalyc.org/journal/3030/303050908010/html", "doi": "10.4025/actasciagron.v39i3.32659", "tier": "A", "quote": "Pichia anomala CCMA0148 and Saccharomyces cerevisiae CCMA0159 provided the greatest inhibition of the growth of all fungal strains.", "accessed": "2026-10-21T00:00:00.000Z" }, { "id": "s5", "title": "Potency of Yeast as a Biocontrol Agent of Ochratoxin A-Producing Fungi and Its Effect on Arabica Coffee Taste", "authors": "Dharmaputra, O. S.; Retnowati, I.; Nurfadila, N.", "year": 2023, "publisher": "BIOTROPIA 30(1), 1–10", "url": "https://journal.biotrop.org/index.php/biotropia/article/view/potency-of-yeast-as-a-biocontrol-agent", "doi": "10.11598/btb.2023.30.1.1379", "tier": "A", "quote": "The results showed that C. krusei (BIO 211287, BIO 211288, and BIO 211289) inhibited A. ochraceus BIO 37310.", "abstractOnly": true, "accessed": "2026-10-21T00:00:00.000Z" }, { "id": "s6", "title": "Inoculation of starter cultures in a semi-dry coffee (Coffea arabica) fermentation process", "authors": "Evangelista, S. R.; Miguel, M. G. C. P.; Cordeiro, C. S.; Silva, C. F.; Pinheiro, A. C. M.; Schwan, R. F.", "year": 2014, "publisher": "Food Microbiology 44, 87–95", "url": "https://www.sciencedirect.com/science/article/pii/S0740002014001191", "doi": "10.1016/j.fm.2014.05.013", "tier": "A", "quote": "DGGE analysis showed that the inoculated yeasts were present throughout the fermentation.", "abstractOnly": true, "accessed": "2026-10-21T00:00:00.000Z" }, { "id": "s7", "title": "Characteristics of fermented coffee inoculated with yeast starter cultures using different inoculation methods", "authors": "Bressani, A. P. P.; Martinez, S. J.; Evangelista, S. R.; Dias, D. R.; Schwan, R. F.", "year": 2018, "publisher": "LWT – Food Science and Technology 92, 212–219", "url": "https://agris.fao.org/search/ru/records/65dfc9cf0f3e94b9e5dc427c", "doi": "10.1016/j.lwt.2018.02.029", "tier": "A", "quote": "Saccharomyces cerevisiae CCMA 0543 inoculated in buckets obtained the highest score in the coffee cupping.", "abstractOnly": true, "accessed": "2026-10-21T00:00:00.000Z" }, { "id": "s8", "title": "Different inoculation methods for semi-dry processed coffee using yeasts as starter cultures", "authors": "Martinez, S. J.; Bressani, A. P. P.; Miguel, M. G. da C. P.; Dias, D. R.; Schwan, R. F.", "year": 2017, "publisher": "Food Research International 102, 333–340", "url": "https://educapes.capes.gov.br/handle/capes/1149793", "doi": "10.1016/j.foodres.2017.09.096", "tier": "A", "quote": "Coffee cupping results proved that both inoculation methods scored well in terms of coffee quality.", "abstractOnly": true, "accessed": "2026-10-21T00:00:00.000Z" }, { "id": "s9", "title": "Coffee protein profiles during fermentation using different yeast inoculation methods", "authors": "Bressani, A. P. P.; Martinez, S. J.; Vilela, L. de F.; Dias, D. R.; Schwan, R. F.", "year": 2020, "publisher": "Pesquisa Agropecuária Brasileira 55", "url": "https://scielo.br/j/pab/a/gkgxHNNPJPJmrNJvn7GFxGj/?lang=en", "doi": "10.1590/s1678-3921.pab2020.v55.01159", "tier": "A", "quote": "left for 16 hours before being transferred to the suspended terraces", "accessed": "2026-10-21T00:00:00.000Z" }, { "id": "s10", "title": "Organic acids produced during fermentation and sensory perception in specialty coffee using yeast starter culture", "authors": "Bressani, A. P. P.; Martinez, S. J.; Sarmento, A. B. I.; Borém, F. M.; Schwan, R. F.", "year": 2020, "publisher": "Food Research International 128, 108773", "url": "https://doi.org/10.1016/j.foodres.2019.108773", "doi": "10.1016/j.foodres.2019.108773", "tier": "A", "quote": "The use of yeasts is an alternative for sensorial differentiation of coffee variety Canario Amarelo and Bourbon Amarelo.", "abstractOnly": true, "accessed": "2026-10-21T00:00:00.000Z" }, { "id": "s11", "title": "Conducting starter culture-controlled fermentations of coffee beans during on-farm wet processing: Growth, metabolic analyses and sensorial effects", "authors": "de Melo Pereira, G. V.; Neto, E.; Thomaz-Soccol, V.; Medeiros, A. B. P.; Woiciechowski, A. L.; Soccol, C. R.", "year": 2015, "publisher": "Food Research International 75, 348–356", "url": "https://agris.fao.org/search/en/records/65df61b963b8185d9cac9b48", "doi": "10.1016/j.foodres.2015.06.027", "tier": "A", "quote": "The inoculation also increased the production of specific volatile aroma compounds (e.g., ethanol, acetaldehyde, ethyl acetate and isoamyl acetate) and decreased the production of lactic acid during the fermentation process.", "abstractOnly": true, "accessed": "2026-10-21T00:00:00.000Z" }, { "id": "s12", "title": "Microbiological and chemical characteristics of wet coffee fermentation inoculated with Hansinaspora uvarum and Pichia kudriavzevii and their impact on coffee sensory quality", "authors": "Elhalis, H.; Cox, J.; Frank, D.; Zhao, J.", "year": 2021, "publisher": "Frontiers in Microbiology 12, 713969", "url": "https://www.frontiersin.org/journals/microbiology/articles/10.3389/fmicb.2021.713969/full", "doi": "10.3389/fmicb.2021.713969", "tier": "A", "quote": "H. uvarum and P. kudriavzevii dominated the fermentations, growing to maximum populations of about 10.0 log CFU/ml compared with 8.0 log CFU/ml in the spontaneous fermentation.", "accessed": "2026-10-21T00:00:00.000Z" }, { "id": "s13", "title": "Effect of Co-Inoculation with Pichia fermentans and Pediococcus acidilactici on Metabolite Produced During Fermentation and Volatile Composition of Coffee Beans", "authors": "da Silva Vale, A.; de Melo Pereira, G. V.; de Carvalho Neto, D. P.; Rodrigues, C.; Pagnoncelli, M. G. B.; Soccol, C. R.", "year": 2019, "publisher": "Fermentation 5(3), 67", "url": "https://doaj.org/article/1b55e10b521b47f9bc8c46d6a61823b0", "doi": "10.3390/fermentation5030067", "tier": "A", "quote": "combined treatment prominently enhanced the aroma complexity of ripe coffee beans", "abstractOnly": true, "accessed": "2026-10-21T00:00:00.000Z" }, { "id": "s14", "title": "Coffee fermentation as a tool for quality improvement: an integrative review and bibliometric analysis", "authors": "dos Santos Silva, M. E.; de Oliveira, R. L.; de Lucena, R. M.; da Silva, S. P.; Porto, T. S.", "year": 2024, "publisher": "International Journal of Food Science and Technology 59(9), 5912–5925", "url": "https://academic.oup.com/ijfst/article/59/9/5912/7911513", "doi": "10.1111/ijfs.17381", "tier": "A", "quote": "all samples, including the control group (without fungal inoculation), obtained a score greater than 85 points.", "accessed": "2026-10-21T00:00:00.000Z" }, { "id": "s15", "title": "Metagenomic, metabolomic, and sensorial characteristics of fermented Coffea arabica L. var. Castillo beans inoculated with microbial starter cultures", "authors": "Madrid-Restrepo, M. A.; León-Inga, A. M.; Peñuela-Martínez, A. E.; Cala, M. P.; Reyes, A.", "year": 2026, "publisher": "mSystems 11(1), e0136425", "url": "https://pubmed.ncbi.nlm.nih.gov/41384736", "doi": "10.1128/msystems.01364-25", "tier": "A", "quote": "Quantitative descriptive analysis revealed that the inoculated coffee outperformed the spontaneous fermentation in all sensory attributes", "abstractOnly": true, "accessed": "2026-10-21T00:00:00.000Z" }, { "id": "s16", "title": "Metagenomic, metabolomic and sensorial characteristics of fermented Coffea arabica L. var. Castillo beans inoculated with microbial starter cultures (preprint)", "authors": "Madrid-Restrepo, M. A.; León-Inga, A. M.; Peñuela-Martínez, A. E.; Cala, M. P.; Reyes, A.", "year": 2025, "publisher": "bioRxiv 2025.01.22.634363", "url": "https://www.biorxiv.org/content/10.1101/2025.01.22.634363v1.full", "doi": "10.1101/2025.01.22.634363", "tier": "B", "tierNote": "Preprint (not peer reviewed) of source 15; cited only for how the trial was set up (starter, added fruit, undisclosed strains), never for a result or a number.", "quote": "Species and strains of added yeast are not detailed due to a verbal confidentiality clause", "accessed": "2026-10-21T00:00:00.000Z" }, { "id": "s17", "title": "Transfer kinetics of labeled aroma compounds from liquid media into coffee beans during simulated wet processing conditions", "authors": "Hadj Salem, F.; Lebrun, M.; Mestres, C.; Sieczkowski, N.; Boulanger, R.; Collignan, A.", "year": 2020, "publisher": "Food Chemistry 322, 126779", "url": "https://agritrop.cirad.fr/595974", "doi": "10.1016/j.foodchem.2020.126779", "tier": "A", "quote": "The parchment resistance significantly affected the mass transfer of the 2-phenylethanol.", "abstractOnly": true, "accessed": "2026-10-21T00:00:00.000Z" }, { "id": "s18", "title": "Mass transfer kinetics of nonvolatile compounds into coffee beans during wet processing: Study at the laboratory scale and in real conditions using two yeast strains", "authors": "Hadj Salem, F.; Vasai, F.; Duez, C.; Sieczkowski, N.; Boulanger, R.; Collignan, A.", "year": 2022, "publisher": "ACS Food Science & Technology 2(5), 852–861", "url": "https://publications.cirad.fr/une_notice.php?dk=601872", "doi": "10.1021/acsfoodscitech.2c00022", "tier": "A", "quote": "Parchment resistance reduced the transfer of these precursors by up to 60%.", "abstractOnly": true, "accessed": "2026-10-21T00:00:00.000Z" }, { "id": "s19", "title": "Modelling the transfer and degradation kinetics of aroma compounds from liquid media into coffee beans during simulated wet processing conditions", "authors": "Hadj Salem, F.; Achir, N.; Sieczkowski, N.; Boulanger, R.; Collignan, A.", "year": 2023, "publisher": "Journal of Food Engineering 343, 111303", "url": "https://agritrop.cirad.fr/603471", "doi": "10.1016/j.jfoodeng.2022.111303", "tier": "A", "quote": "The transfer of 2-phenylethanol was well described by a model including evolving resistance over time (R2 = 0.98)", "abstractOnly": true, "accessed": "2026-10-21T00:00:00.000Z" }, { "id": "s20", "title": "Effect of solid-state fungal fermentation on the non-volatiles content and volatiles composition of Coffea canephora (Robusta) coffee beans", "authors": "Tang, V. C. Y.; Sun, J.; Cornuz, M.; Yu, B.; Lassabliere, B.", "year": 2021, "publisher": "Food Chemistry 337, 128023", "url": "https://doi.org/10.1016/j.foodchem.2020.128023", "doi": "10.1016/j.foodchem.2020.128023", "tier": "A", "quote": "Significant strain-specific differences were observed in the contents of aroma compounds after roasting.", "abstractOnly": true, "accessed": "2026-10-21T00:00:00.000Z" }, { "id": "s21", "title": "2026 World Brewers Cup Official Rules and Regulations", "authors": "WCC Rules and Regulations Committee", "year": 2026, "publisher": "World Coffee Championships (Specialty Coffee Association)", "url": "https://wcc.coffee/s/2026-World-Brewers-Cup-Rules-and-Regulations.pdf", "tier": "B", "quote": "For the purposes of this competition, no additives of any kind may be added to coffee after it reaches the \"green coffee\" stage", "accessed": "2026-10-21T00:00:00.000Z" }, { "id": "s22", "title": "Brazil 2023 – Cup of Excellence", "authors": "Alliance for Coffee Excellence", "year": 2023, "publisher": "Alliance for Coffee Excellence", "url": "https://allianceforcoffeeexcellence.org/brazil-2023/", "tier": "B", "quote": "ensuring no additives other than water are introduced", "accessed": "2026-10-21T00:00:00.000Z" }, { "id": "s23", "title": "Influence of anaerobic fermentation and yeast inoculation on the viability, chemical composition, and quality of coffee", "authors": "Jimenez, E. J. M.; Martins, P. M. M.; Vilela, A. L. de O.; Batista, N. N.; da Rosa, S. D. V. F.; Dias, D. R.; Schwan, R. F.", "year": 2023, "publisher": "Food Bioscience 51, 102218", "url": "https://www.alice.cnptia.embrapa.br/alice/bitstream/doc/1148954/1/Influence-of-anaerobic-fermentation-and-yeast-inoculation-on-the-viability.pdf", "doi": "10.1016/j.fbio.2022.102218", "tier": "A", "quote": "The T. delbrueckii yeast obtained the best score of 86.50 for pulped coffee", "accessed": "2026-10-21T00:00:00.000Z" }, { "id": "s24", "title": "Exploring the impacts of postharvest processing on the microbiota and metabolite profiles during green coffee bean production", "authors": "De Bruyn, F.; Zhang, S. J.; Pothakos, V.; Torres, J.; Lambot, C.; Moroni, A. V.; Callanan, M.; Sybesma, W.; Weckx, S.; De Vuyst, L.", "year": 2017, "publisher": "Applied and Environmental Microbiology 83(1), e02398-16", "url": "https://journals.asm.org/doi/10.1128/aem.02398-16", "doi": "10.1128/AEM.02398-16", "tier": "A", "quote": "Several of the metabolites associated with lactic acid bacterial metabolism (e.g., lactic acid, acetic acid, and mannitol) produced in the mucilage were also found in the endosperm.", "accessed": "2026-10-21T00:00:00.000Z" } ], "markdown": "## Two meanings of co-fermentation\n\nAfter harvest, microbes on the fruit and in the processing environment ferment the coffee.24 Producers who want more control add something to this fermentation.1 This entry uses the word “co-fermentation” for two different additions.\n\n1. **Starter cultures.** A selected strain of yeast or lactic acid bacteria is added to the cherries or the fermentation tank. Researchers call this inoculation.7,11 When yeasts and bacteria are added together, studies speak of co-inoculation.13\n2. **Added ingredients.** Fruit, spices, juices or other materials are fermented together with the coffee. This entry calls such lots “infused”. In one Colombian trial, crushed pineapple went into the tank together with a yeast starter.16\n\nThe two practices can be combined, as in that trial. That makes it hard to say which addition caused a change in the cup.16 How spontaneous fermentation works is covered under fermentation; the microbes themselves under microorganisms.\n\n## What a starter culture is meant to do\n\nA starter culture is a large dose of one or more chosen strains. In one laboratory trial, the added yeasts grew to about 100 times the yeast population of a spontaneous fermentation.12 The aim is that this strain leads the fermentation instead of whatever microbes happen to be present.\n\nThree goals recur in the studies:\n\n- **Faster, more complete mucilage breakdown.** Many yeasts and bacteria make enzymes that break down pectin, the gel that holds the mucilage together.1\n- **Aroma.** Starter yeasts raised compounds such as ethyl acetate and isoamyl acetate, which smell fruity.11 In one trial, only the inoculated coffee showed a caramel flavour.6\n- **Protection.** Some yeasts slow the growth of moulds that produce ochratoxin A, a toxin.4,5\n\nStarters have been tested in natural, semi-dry (pulped natural) and washed coffee, and in closed-tank fermentation.6,7,11,23\n\n## How a strain is chosen\n\nMany candidate strains come from coffee itself. Researchers isolate yeasts from coffee fermentations and test them one by one.3,4 A 2021 review lists the main criteria: pectin-degrading enzymes, production of aroma compounds, the chemical changes the strain causes, and no unwanted metabolites.2\n\n
\n
\n
How candidate starter strains are narrowed down in the studies of this entry: isolation, enzyme screens, tests against toxin-producing moulds, then trials on coffee. Schematic; no single study ran every step.2,3,4,7
\n
\n\nLaboratory screens come first. In a 2019 study, eight of 28 yeast isolates showed pectin-degrading activity on test plates.3 In a 2017 study, 32 wild yeasts were grown next to toxin-producing moulds to see which held them back.4 A strain that passes is then tried on coffee. There it must survive and stay present until drying; trials check this with DNA-based methods.6,7\n\n## Fermenting with fruit, spices and other ingredients\n\nIn the second meaning, producers add other materials to the fermentation. For this practice, the peer-reviewed record is thin. No study opened for this entry compares the same coffee fermented with and without added fruit alone.\n\nWhat is known comes from two directions. First, laboratory work shows that aroma compounds dissolved in the fermentation liquid can move into depulped seeds.17 The parchment slows this down, and some compounds are broken down on the way.17,18 Second, the one trial with added fruit also used a yeast starter and other ingredients, so their effects cannot be separated.16\n\n## What the rules allow\n\nThe World Brewers Cup draws its line at the green coffee stage. Its 2026 rules allow no additives of any kind once coffee has become green coffee.21 Additions during fermentation, before that stage, are therefore not excluded by this wording.21 One Cup of Excellence programme set a stricter line: its 2023 Experimentals category in Brazil admitted no additives other than water.22\n\n
\n
\n
Where starter cultures, added ingredients and fungal fermentation of green coffee enter the chain, and where the 2026 World Brewers Cup rule draws its line. Schematic, not to scale.7,9,16,20,21,22
\n
\n\n\n\n## Inoculation in practice\n\n**On cherries for natural and semi-dry coffee.** A Brazilian research group tested two methods. In the direct method, a yeast suspension was sprayed onto cherries already spread on raised drying terraces. In the bucket method, the suspension was mixed with the cherries in a bucket and left for 16 hours before drying.7,9 Cells from a 500 mL suspension were applied to 10 kg of coffee, giving 10⁷ cells/g.9 The bucket method favoured the persistence of the starters, especially *Saccharomyces cerevisiae*, in semi-dry coffee.8\n\n**In wet and closed-tank fermentation.** In an on-farm wet trial, *Pichia fermentans* YC5.2 was inoculated with or without 2% (w/v) sucrose11; a 2024 review gives the inoculum as 10⁷ cells/mL.14 In self-induced anaerobic fermentation (SIAF) in 50 L closed bioreactors, three yeasts were added at 10⁷–10⁸ cells/mL.23\n\n**Checking that the starter took hold.** Persistence has been followed by DGGE fingerprinting and by quantitative PCR (qPCR).6,7,8 In the semi-dry trial, the inoculated yeasts were present throughout fermentation.6 Whether a starter displaces the resident bacteria differs between studies; this is covered under microorganisms.\n\n## Reference protocol\n\nThe table brings together the parameters reported in the trials. They are research settings, not a commercial recipe.\n\n| Parameter | Value reported | Source |\n|---|---|---|\n| Strains in natural and semi-dry trials | *S. cerevisiae* CCMA 0543, *Candida parapsilosis* CCMA 0544, *Torulaspora delbrueckii* CCMA 0684; *Meyerozyma caribbica* CCMA 0198 | Bressani et al. 20187, Martinez et al.8, Bressani et al. 202010 |\n| Strains in wet trials | *Pichia fermentans* YC5.2; *Hanseniaspora uvarum*, *Pichia kudriavzevii*; *P. fermentans* with *Pediococcus acidilactici* | de Melo Pereira et al.11, Elhalis et al.12, da Silva Vale et al.13 |\n| Dose on cherries | 10⁷ cells/g (500 mL suspension on 10 kg) | Bressani et al. 20209 |\n| Dose in liquid | 10⁷ cells/mL, with or without 2% (w/v) sucrose (wet); 10⁷–10⁸ cells/mL (SIAF) | de Melo Pereira et al.11, dos Santos Silva et al.14, Jimenez et al.23 |\n| Application on cherries | Sprayed on the drying terrace, or mixed in a bucket and held 16 h before drying | Bressani et al.7,9 |\n| Peak yeast population | About 10.0 log CFU/mL inoculated vs 8.0 log CFU/mL spontaneous (wet, laboratory scale) | Elhalis et al.12 |\n| Persistence check | DGGE; qPCR | Evangelista et al.6, Bressani et al.7 |\n| Pectinolytic screen | 8 of 28 isolates positive; pectin degradation index 110–178% | Haile & Kang3 |\n| Antagonism screen (in vitro) | 32 isolates co-cultured with *Aspergillus*, 28 °C, 7 days | de Souza et al.4 |\n| Competition limit | No additives after the green coffee stage | World Brewers Cup 202621 |\n\n## Selecting strains: the evidence\n\n**Pectin-degrading enzymes.** Pectinolytic activity has been found in coffee isolates of the bacteria *Klebsiella*, *Erwinia*, *Bacillus* and *Lactobacillus*, and of the yeasts *Pichia*, *Saccharomyces* and *Candida*.1 In a 2019 screen, 28 yeasts were isolated and eight showed pectin-degrading activity; their pectin degradation index ranged from 110 to 178%, highest in *Saccharomycopsis fibuligera* KNU18Y4.3 The same study measured polygalacturonase, pectin lyase and pectin methylesterase; *Wickerhamomyces anomalus* KNU18Y3 reached 16.96 U/mL pectin lyase.3\n\n**Aroma and absence of off-products.** A 2021 review names volatile production and the lack of undesirable metabolites as selection criteria, beside pectin lyase secretion.2 In the semi-dry trial, butyric and propionic acids were absent from all treatments.6 These acids can give an onion flavour.1\n\n**Adaptation.** The same review describes *S. cerevisiae* strains as showing good adaptability to the fermentation process.2 No study opened for this entry tested candidate strains for tolerance of temperature, pH or ethanol before field use; stress tolerance is therefore not documented here as a measured criterion.\n\n**Antagonism against ochratoxin A moulds.** In a 2017 laboratory study, 32 wild yeasts from coffee and cocoa fermentations were co-cultured with *Aspergillus carbonarius* and *A. ochraceus* at 28 °C for seven days.4 *Pichia anomala* CCMA0148 and *S. cerevisiae* CCMA0159 inhibited growth of all fungal strains most.4 The abstract reports that *Rhodotorula mucilaginosa* inhibited ochratoxin A production by all three isolates, but the results also list *R. mucilaginosa* isolates that stimulated it in some tests.4 These were culture-media tests, and the authors call for tests on coffee fruit.4 A 2023 Indonesian study went further: three *Candida krusei* isolates inhibited *A. ochraceus*, grew on wet- and semi-wet-processed beans, and gave higher total cupping scores than uninoculated beans and beans treated with a commercial lactic acid bacteria inoculant.5\n\n## What trials found: chemistry and the cup\n\n**Acids and volatiles.** In natural coffee, all directly inoculated treatments had high citric and malic acid, and 217 volatile compounds were identified.7 After roasting, caramel-related volatiles appeared with *S. cerevisiae* and fruity (apple, cherry) compounds with *C. parapsilosis*.7 In wet fermentation, *P. fermentans* raised ethanol, acetaldehyde, ethyl acetate and isoamyl acetate and lowered lactic acid.11 In semi-dry coffee, acids, pyrazines and pyridines were the main volatiles in green and roasted beans.8\n\n**Strain, method and variety interact.** *S. cerevisiae* CCMA 0543 scored highest when inoculated in buckets; *C. parapsilosis* did better sprayed directly; *T. delbrueckii* did well with both methods.7 In a 2020 trial with four starters, two varieties and two processing methods, the perceived fruity, nutty and cocoa attributes varied with variety, processing and inoculum.10 A strain that works for one coffee and method cannot be assumed to work for another.7,10\n\n**Co-inoculation.** Combining *P. fermentans* with the lactic acid bacterium *Pediococcus acidilactici* on ripe cherries raised sugar consumption and the production of lactic acid, ethanol and ethyl acetate; the authors suggest a synergy between the two.13 With immature cherries, only *P. fermentans* alone improved metabolite formation, and the combined treatment had no significant effect.13\n\n**Cup scores.** Results are mixed (see Contested). In closed-tank trials, *T. delbrueckii* gave 86.50 points for pulped coffee and *C. parapsilosis* 85.90 for natural coffee, against 84.25 and 83.75 for conventional coffee.23 These figures compare inoculated closed-tank coffee with conventional coffee, so the vessel and the yeast changed together.23 In a laboratory-scale wet trial, inoculated coffees scored higher for flavour, aroma and acidity, but differences in sweetness, balance, clean cup, uniformity, aftertaste and overall score were reported as not statistically significant.12 The same paper marks the highest overall score (82.0, mixed yeasts) with p < 0.05, so its reporting of the overall score is not consistent.12 A 2024 review notes a yeast trial in which every sample, including the uninoculated control, scored above 85 points.14\n\n**The Colombian starter-plus-fruit trial.** In a 2026 study of Castillo coffee, the inoculated wet fermentation outperformed spontaneous fermentation in all sensory attributes by quantitative descriptive analysis, with over a hundred differential metabolites.15 Inoculated coffee had more sucrose, mannitol, methyl phenylacetate, malic, citric and quinic acid.15 In the preprint, the starter was a commercial yeast cocktail with flour and a sugar source, added with crushed pineapple; the strains were confidential.16\n\n## How compounds reach the seed\n\nMicrobial metabolites made in the mucilage – lactic acid, acetic acid and mannitol – have been found in the endosperm of spontaneously fermented coffee.24 A French research group measured the transfer directly, using labelled compounds in simulated wet processing.\n\n**Volatiles.** Labelled butanal, 2-phenylethanol and isoamyl acetate were added to the liquid at 25 °C and sampled from 0 to 48 hours.17 After 12 hours in the medium with depulped beans and yeast, the beans held 0.2 ± 0.03 µg/g butanal, 11.2 ± 0.66 µg/g 2-phenylethanol and 1.3 ± 0.04 µg/g isoamyl acetate.17 The parchment slowed 2-phenylethanol significantly; butanal and isoamyl acetate were partly metabolised; and the yeast reduced butanal transfer.17\n\n
\n
\n
Labelled aroma compounds found in depulped coffee seeds after 12 hours in a simulated wet fermentation with yeast at 25 °C (mean ± spread as given in the abstract). The three compounds entered at very different rates.17
\n
\n\n**Non-volatile precursors.** Lactic acid, alanine and glutamic acid also moved into the beans; parchment reduced the transfer of these precursors by up to 60%.18 In real fermentations with two *S. cerevisiae* strains, lactic acid and alanine were produced and entered the beans after a 12-hour lag.18 Fructose underwent germination-related metabolic reactions inside the seed.18\n\n**Models.** A 2023 model described 2-phenylethanol transfer well with a resistance that changes over time (R² = 0.98); butanal and isoamyl acetate fitted better with two parallel first-order reactions.19\n\nThese studies used single compounds in solution, not fruit. They show that dissolved aroma compounds can enter the seed. They do not show how much of a fruit's aroma survives roasting.17,19\n\n## Fermenting green coffee\n\nA third practice ferments the green coffee itself, after drying and hulling. In solid-state fermentation of green canephora with *Aspergillus* and *Mucor* species, free sugars, free amino acids and caffeoylquinic acids changed before roasting.20 After roasting, *A. oryzae* raised pyrazines and *Mucor plumbeus* raised furans.20 This treatment falls after the green coffee stage, where the 2026 World Brewers Cup rules allow no additives.21\n\n## Competition rules and disclosure\n\nThe 2026 World Brewers Cup rules define green coffee as seeds of the genus *Coffea*, dried in post-harvest processing and free of all pericarp layers.21 After that stage, no additives of any kind may be added, including exposure to aromatic substances, flavourings, perfumes, liquids or powders.21 The brewed beverage may contain no additives beyond whole-bean coffee, brew water, brewing devices and service vessels.21\n\nThe Cup of Excellence set a narrower boundary in Brazil in 2023: its Experimentals category included “both wet and dry processes with unique treatments”, with no additives other than water.22\n\nNo standard opened for this entry requires a producer or seller to declare added fruit, spices or starter cultures on a green or roasted coffee lot.\n\n## Risks and limits\n\n- **Confounded comparisons.** Several trials change more than one thing at once: the vessel and the yeast,23 or the yeast, flour, sugar and fruit.16 The effect of the starter alone is then unknown.\n- **Non-significant score gains.** In some trials, higher mean scores were not statistically significant.12\n- **Fruit quality matters.** Mixed cultures that helped ripe cherries did not help immature ones.13\n- **Laboratory versus farm.** Antagonism against ochratoxin A moulds was shown mainly on culture media; evidence on coffee seeds is limited to one study opened for this entry.4,5\n- **Unknown strains.** Commercial cocktails may not name their strains, which prevents replication.16\n\n## Open questions\n\n- **Fruit volatiles in the cup.** No study opened for this entry traced compounds from added fruit or spices into roasted coffee.\n- **Detection.** No validated method for telling an infused lot from an uninfused one was found among the sources of this entry.\n- **Food labelling.** No food-law source opened for this entry addresses co-fermented coffee; its labelling status is not established here.\n- **Stress tolerance.** Tolerance of heat, acidity and ethanol is not reported as a screening step in the studies opened.2,3\n- **Which microbes do the work.** The authors of the 2026 Colombian study call for separating the contributions of each microbe and pathway.15\n\n## Key studies\n\n
\n

Bressani et al., 2018 – Spray or bucket. Inoculates natural coffee with S. cerevisiae, C. parapsilosis or T. delbrueckii, either sprayed on the drying terrace or held 16 hours in buckets. Each strain responded differently to the method; S. cerevisiae in buckets scored best, and inoculation shifted acids and roasted volatiles.

\n

LWT – Food Science and Technology 92, 212–219 · doi:10.1016/j.lwt.2018.02.029

\n
\n\n
\n

de Souza et al., 2017 – Yeasts against ochratoxigenic moulds. Tests 32 wild yeasts from coffee and cocoa against three Aspergillus isolates on culture media. Several Pichia and Saccharomyces strains held back fungal growth and sporulation; effects on ochratoxin A production differed between isolates and conditions.

\n

Acta Scientiarum. Agronomy 39(3), 349–358 · doi:10.4025/actasciagron.v39i3.32659

\n
\n\n
\n

da Silva Vale et al., 2019 – Yeast plus lactic acid bacteria. Compares single and combined inoculation with Pichia fermentans and Pediococcus acidilactici on ripe and immature fruit. The pair worked together on ripe fruit, raising acids, alcohols and aroma complexity, but not on immature fruit.

\n

Fermentation 5(3), 67 · doi:10.3390/fermentation5030067

\n
\n\n
\n

Hadj Salem et al., 2020 – Do aroma compounds enter the seed? Follows labelled butanal, 2-phenylethanol and isoamyl acetate from a liquid into coffee seeds under simulated wet processing. All three entered at different rates; the parchment slowed 2-phenylethanol, and the others were partly metabolised.

\n

Food Chemistry 322, 126779 · doi:10.1016/j.foodchem.2020.126779

\n
\n\n
\n

Madrid-Restrepo et al., 2026 – A commercial starter in Colombia. Compares an inoculated and a spontaneous wet fermentation of Castillo coffee with sequencing, metabolomics and sensory panels. The inoculated coffee rated higher and differed in over a hundred metabolites, but fruit and other ingredients were added with the starter.

\n

mSystems 11(1), e0136425 · doi:10.1128/msystems.01364-25

\n
" }