Figure: Dr. Guo Shoujie, the initiator of Tianfeng Intelligence, and his team
Tianfeng Intelligence is a fusion application of "regular breeding + biotechnology + information technology + artificial intelligence " in the technical route. The "seed accelerator" is an artificial intelligence breeding service platform based on machine learning (Machine Learning). But in terms of company name and product introduction, they are very pragmatic. This can also be seen from the recruitment information for digital breeding engineer positions previously released by the company:
What is the truth of biological breeding? is a unique genius gene catcher with keen eyes? Is it a lot of brains to play with various and ? Or is it willing to hide behind the scenes, change one's fate and create a new species... Perhaps all this is both the truth and the whole illusion. We decided not to talk about feelings, but to tell the truth: breeding research, even with fashionable AI, has the dilemma that any work is inevitable, just like the confusion that our previous generation of breeders hopes to find good materials that can stably inherit in a vast resource, and we also try our best to make this process efficient. We also firmly believe that only by knowing the direction can we walk far away; only by bending can we pick up wheat ears. All subversion occurs after the inheritance that respects science, tradition, and stands on the shoulders of giants. |
"Seed Accelerator" is based on the scientific model shown in the figure below. It learns the genetic data and phenotypic data of existing varieties through a variety of optimized machine learning algorithms (at the same time, the system supports grouping of germplasm data, improving the consistency of germplasm to improve prediction accuracy), and predicts breeding materials based on the learning results, helping breeders choose more advantageous breeding materials for trial seeding, screening, and combining, and improving breeding efficiency.
Figure: A scientific model of whole genome selection breeding based on machine learning
Tianfeng Wisdom's product and service characteristics are mainly reflected in the following aspects. First of all, in terms of ease of use: the seed accelerator platform is an intuitive and easy-to-use digital breeding platform. There is no need for the breeding teacher to design and use complex programming and data processing by himself, whether it is traditional field data collection or various phenotypic data collected by drone , spectroscopy and other equipment, through the association with the automated processing of genotype data, and then use the TFseeds embedded platform to train multiple AI breeding models, you can give prediction and selection schemes (GS) in a "plug-and-play" manner, and provide various screening functions and final breeding scheme downloads to achieve the role of assisting breeders to efficiently select breeding materials, reduce trial cost, and improve breeding efficiency.
Tianfeng Seed Accelerator Platform
Secondly, from the perspective of crop categories: Tianfeng Wisdom did not select corn, fruits and vegetables and other categories with the largest market size, but focused on barley, soybean and other directions with greater opportunities in the industrial system, making full use of the advantages of domestic germplasm resources and foreign mature breeding selection models, and actively promote the construction of digital agriculture, i.e., the ecosystem led by the seed industry. For example, in the field of non-GMO soybean , the company has jointly carried out an industrial cooperation model of "from seeds to processing, from seeds to tables" with the Harbin listed company Oriental Group soybean planting base, and has carried out breeding research and development and industrial promotion for special beans. In the barley field, we use the advantages of digital breeding platforms to develop special varieties for plant factories and large-scale beer planting and other ways to organize cooperation in the industrial chain.
Figure: Real scene of molecular detection laboratories and wheatgrass factories
Furthermore, with the maturity of molecular identification technology and the sharp decline in costs, the accumulation of phenotype and environmental data is increasing, and digital means to assist in breeding more and more leading seed companies such as Bayer , Codiva, etc. as the basis for commercial breeding. from the process , commercial breeding emphasizes team combat and resource sharing, decomposing all links of breeding, but it can efficiently and smoothly flow, truly realizing "large-scale, high-throughput, standardized" breeding. Tianfeng Smart’s AI breeding platform is based on the general background of the process of commercial breeding system in China’s seed industry. It not only technically but also processally, it assists various traditional breeding teams in China to achieve smooth digitalization and intelligent upgrades through the seed accelerator platform.
Tianfeng's "Seed Accelerator Platform" uses a large amount of publicly published data and foreign data, and uses a variety of artificial intelligence models to improve the effectiveness of predictions, and has been verified by some domestic breeders. But at the same time, models that apply different breeding strategies, real, accurate, complete and comprehensive data are the basis for the continuous improvement of the effectiveness of digital breeding models. The accumulation of data, because of the irreversibility of time and space, must be the core competitiveness of each breeding team. In order to break the limitations and barriers of this data, although Tianfeng Wisdom is a highly focused small team, it still positions its vision as: uses the internationally leading digital breeding platform to promote the commercial breeding process of China's seed industry with domestic breeders. looks forward to more breeding teams reaching cooperation with Tianfeng Wisdom.
Figure: Breeding industry chain partners exchange
1. China has broad prospects for commercial breeding, and there are conditions to support the birth of "big aircraft carrier"
China has more than 7,000 seed companies, and the world's largest but most scattered breeding team as the main form. It has given AI breeding digital services a large enough market capacity. It is expected that by 2027, the global market in the digital breeding field will exceed 140 billion yuan. In the 14th Five-Year Plan, the construction of a digital breeding platform and the realization of intelligent breeding and precise breeding will be the key research directions. However, my country's seed companies are relatively small in scale, and currently more than 500 million can rank among the top ten in China. According to the maximum R&D funding of 4%, it is estimated that a Chinese seed company entering the top ten has less than 20 million in annual R&D expenses, which is not enough to support the R&D team with a background in biological genetics, , information technology, and computer complexity in the era of big data. This has also made my country delayed the real implementation of digital breeding.
On the other hand, the arrival of the era of precision breeding + digital agriculture has given order agriculture and the business model from seed to processing and from seed to dining table. There is a new planting market of 10 billion yuan in the soybean field alone. From the perspective of the protein market, there is a high-energy plant protein food market of hundreds of billions of yuan in the domestic non-GMO soybean industry. Large-scale order agriculture, the "variety + brand" model, can obviously provide protection capabilities beyond traceability.
Figure: Tianfeng Seed Accelerator Platform Data
As a weapon for precise breeding, the digital breeding platform can comprehensively consider the different needs of consumers, growers, processing plants, and food companies. For growers: short plant, dense planting, mechanization, weeding, insect resistance, etc., it is more urgent; from the processing end: If the process of processing protein changes from three to two, the cost of protein companies will be greatly reduced without changing other conditions; the demand on the consumer market is more diversified, and traditional soy sauce, tofu, soy milk, bean curd, bean curd, bean paste, etc. have different focuses, artificial meat, functional food, snacks and snacks, infant supplementary food , etc. have also proposed new directions.
Commercial breeding is different from scientific research breeding. Scientific research focuses on finding gene switches and making breakthroughs in single traits. These have made particularly great contributions to the development of biology. However, successful large varieties in commercial areas often focus not only on maximizing yields, but also on a series of traits based on the needs of different natural environments: comprehensive cold and drought resistance, high temperature and water resistance, disease, insects and salt and alkali resistance, but these traits are often not controlled by single genes, but are determined by multiple genes and interactions with environmental factors. It is difficult to achieve comprehensive testing of different traits at the same time using traditional field experiment methods.Therefore, the future application prospects for the use of the integrated application of "regular breeding + biotechnology + information technology + artificial intelligence" are broad.
2. Investors say
Ferry Theory | Chiaxin Fund's commitment to bioagriculture
Chiaxin Fund has always focused on early investment and incubation in the intersection of bioagriculture and data science. "Joint entrepreneurship and industrial empowerment" is a concept that Chia Xin has always adhered to. As an industrial fund in a sub-track, it has successfully invested and incubated China's largest field crops such as wheat and cotton, biological breeding company, Jiushenghe Seed Industry and and Compson Bio, the leader in animal genome breeding. These two startups have been successfully selected as Chinese seed industry formation companies and become the vanguard of China's ideal of a strong seed industry country.
was introduced to Xinjiang from Compson Bio, a pure technology company, transforming from a full-chain biotechnology company that has developed in-depth in the fields of breeding sheep, cattle, , etc., to a certain extent, verified the keen insight and execution ability of Caixin Fund in the direction of technological innovation and industrial transformation.
In 2022, Charles Schwab Fund continued to practice the ferry theory in the field of seed technology. In addition to the traditional investment "financial", Charles Schwab Fund hopes to build Tianfeng Wisdom into a technology output platform "technology" that empowers other seed companies. The pioneering attempts of grass-fed and livestock integration such as "Xincao Company" have also made the incubation model of Charles Schwab Entrepreneurship Accelerator more mature, making Charles Schwab Fund itself an institution for "people" output.
