- Cross-validation of a semantic segmentation network for natural history collection specimens. Computers can distinguish the herbarium label from the actual specimen and other stuff on the sheet, helping with the whole automatic digitization thing, but it takes some really fancy math.
- The landscapes of livestock diversity: grazing local breeds as a proxy for domesticated species adaptation to the environment. Medium fancy math used to map breed diversity in the Iberian Peninsula for different livestock species and relate it to environmental factors.
- South-to-north migration preceded the advent of intensive farming in the Maya region. Sort of like tomato, but in the other direction. Plenty of math involved, but behind the scenes, thankfully.
- Timing and magnitude of climate-driven range shifts in transboundary fish stocks challenge their management. Huge amount of data and very fancy math shows fish are in trouble.
- A Core Set of Snap Bean Genotypes Established by Phenotyping a Large Panel Collected in Europe. Ok, even I can follow the math on this one.
- Assessment of biogeographic variation in traits of Lewis flax (Linum lewisii) for use in restoration and agriculture. Very approachable math shows which populations of a CWR can best be used for restoration, and where; and also for domestication and breeding.
- An updated checklist of plant agrobiodiversity of northern Italy. Very useful use of very basic maths. Key number: only 43% of the PGR on the list are conserved ex situ.
- Cannabis, the multibillion dollar plant that no genebank wanted. No math needed to figure out weed needs a genebank.
A new genebank for the ages is set for ages
Great news from the opening ceremony of the new Future Seeds genebank in Palmira, Colombia on 15 March:
The Bezos Earth Fund pledged US$17 million for Future Seeds, a new CGIAR genebank inaugurated today. The new genebank will bolster global efforts to safeguard the world’s future food supply.
This genebank is truly next-level:
Future Seeds is the most advanced facility in Latin America and is expected to become the first ever platinum-level LEED (Leadership in Energy and Environmental Design)-certified genebank building in the world. Its Data Discovery and Biotechnology Lab will use big-data technologies to mine the genebank using the latest in genetics to document the range of possibly useful traits in the current collection. Other breakthrough technologies across genebanks include drones and robotic rovers, which are helping analyze crop characteristics in the field more rapidly, and the use of artificial intelligence to enable collectors to identify potential biodiversity hotspots in nature.
Here, check it out for yourselves:
And here’s an overview of the collections from Genesys (beans in red, cassava blue, forages green).
Full disclosure: we also support the place at work.
Cropland *could* be almost halved
I’m recycling this from Jeremy’s latest newsletter, with permission. So I don’t have to write something on the paper in question myself, as I originally planned.
I’m honestly not sure what to make of this recent paper: Global cropland could be almost halved: Assessment of land saving potentials under different strategies and implications for agricultural markets. The gist of it seems to be that if we were able to grow crops more productively (closing the yield gap, as it is known) we would need less land, reduce crop prices, and cure the common cold.
Not quite, obviously, but this kind of model-based approach to transforming global agriculture seems to me to be long on possibilities and short on practicalities. Of course, the modellers could point out that they are merely showing the way and that others will have to make the decision to take us down the road. Points, too, for figuring out how all this might affect prices and global trade flows. However, I remain befuddled and bemused, as I was when I first encountered this sort of study in 2009 and then again in 2017.
Brainfood: Spatial data, Extinction risk, Improved lentils, Lentil collection, Ohia germination, Shea genomics, Wild olive, Cacao climate refugia, Cacao sacred groves, Italian winter squash, Nigerian yams, Bambara groundnut diversity
- CropHarvest: A global dataset for crop-type classification. 90,000 datapoints all over the world, nicely labelled with what’s going on there agriculturally speaking. Let the AI rip.
- Using publicly available data to conduct rapid assessments of extinction risk. Pretty much useless, but at least now we know why. Should have used AI.
- Plot-level impacts of improved lentil varieties in Bangladesh. About 15% higher yields and gross margins, resulting in lots of savings on imports.
- Agro-Morphological Characterization of Lentil Germplasm of Indian National Genebank and Development of a Core Set for Efficient Utilization in Lentil Improvement Programs. And a core subset to boot. Unclear if any were used to breed the above.
- Variation in Germination Traits Inform Conservation Planning of Hawaiʻi’s Foundational ʻŌhiʻa Trees. Germination was lower from some populations than from others, but not because of environmental factors.
- Genomic Resources to Guide Improvement of the Shea Tree. Ok, great, but now what exactly? And no word on germination…
- Current Status of Biodiversity Assessment and Conservation of Wild Olive (Olea europaea L. subsp. europaea var. sylvestris). When can we expect something similar for shea tree?
- Extreme climate refugia: a case study of wild relatives of cacao (Theobroma cacao) in Colombia. The forest areas where wild cacao has survived the longest, and is particularly diverse, will be cut in half in 50 years. I wonder what the figures are for wild olive.
- Soil biomarkers of cacao tree cultivation in the sacred cacao groves of the northern Maya lowlands. Maybe re-introduce it? More here.
- How to save a landrace from extinction: the example of a winter squash landrace (Cucurbita maxima Duchesne) in Northern Italy (Lungavilla-Pavia). It’s great to have ‘Berrettina di Lungavilla’ back, but 7 years for one landrace? No sacred groves involved. Shea harvesters unavailable for comment.
- Collection, characterizaton, product quality evaluation, and conservation of genetic resources of yam (Dioscorea spp.) cultivars from Ekiti State, Nigeria. At least it’s more than one landrace.
- Genetic Diversity and Environmental Influence on Growth and Yield Parameters of Bambara Groundnut. 95 landraces, no less. All safe from extinction. Right?
Brainfood: Aspen mapping, Biodiversity & ag, Mining forages, China forages, China groundnuts, Soil microbes, Agroecology messaging, Old wood, Ugandan sorghum, New wild sweetpotato, Tasty fruits
- Remote sensing of cytotype and its consequences for canopy damage in quaking aspen. You can tell diploid from triploid trees from space.
- Future global conflict risk hotspots between biodiversity conservation and food security: 10 countries and 7 Biodiversity Hotspots. Fancy maths tells us biodiversity and agriculture are most in conflict in DRC, Sierra Leone, Malawi, Togo, Zambia, Angola, Guinea, Nigeria, Laos, and Cambodia.
- Allele mining in diverse accessions of tropical grasses to improve forage quality and reduce environmental impact. A draft reference genome from a single species tells us about 7 potentially useful alleles among 104 clearly very well chosen accessions of Urochloa spp and Megathyrsus maximus.
- Research Status of Forage Seed Industry in China. I wonder how many of the above alleles can be found in the Chinese forage collection. Might be easier to eventually find out if the website supposedly serving up the national forage germplasm resource management system actually worked.
- Safe conservation and utilization of peanut germplasm resources in the Oil Crops Middle-term Genebank of China. We are even told about some individual interesting accessions, though not how to get hold of them.
- The impact of crop diversification, tillage and fertilization type on soil total microbial, fungal and bacterial abundance: A worldwide meta-analysis of agricultural sites. Meta-analysis tells us that use of organic fertilisers and reduced tillage are associated with more microbes, fungi and bacteria in the soil.
- Detecting the linkage between arable land use and poverty using machine learning methods at global perspective. Machines tells us that higher crop yields and more fertilisers are associated with lower poverty levels. Non-machines are shocked. No word on soil microbial abundance.
- The 10 Elements of Agroecology: enabling transitions towards sustainable agriculture and food systems through visual narratives. Well, these 10 are not only the elements of agroecology, so they could tell us about other messaging too.
- Regional Patterns of Late Medieval and Early Modern European Building Activity Revealed by Felling Dates. Tree rings in old buildings tells us more felling where and when grain prices were low and mining activity high. No machines involed.
- Genetic diversity analysis and characterization of Ugandan sorghum. A tropical genebank collection can tell us about temperate-adapted germplasm, if we know how to ask.
- Discovery and characterization of sweetpotato’s closest tetraploid relative. Meet Ipomoea aequatoriensis T. Wells & P. Muñoz sp. nov. from, well, Ecuador.
- Metabolomic selection for enhanced fruit flavor. Another machine tells us how to pick tasty tomatoes and blueberries from chemical profiles. No word on when it will be able to describe new species.