AGROCHEMICAL 

Applying cheminformatics to agricultural data gives rise to useful insights within the agrochemical industry. Computational techniques used for the collection, management and analysis of agricultural data allows for the optimisation of processes and design of useful software/equipment within the industry.

Implementation of artificial intelligence (AI) to agro-informatics allows for accurate analysis and management of data such as satellite imagery, weather data, animal disease monitoring, to help improve the sustainability and economics of agriculture practises such as farming.

Agro informatics has important applications, particularly in areas such as food production and sustainability.

FLAVOUR / FRAGRANCE 

Flavour and Fragrance cheminformatics combines scientific data with consumer preferences, looking at datasets that predict trends in flavours and fragrances. Informatics in the flavour and fragrance industries provides confidence in data-guided decision making where products are often very chemically complex. The efficient formulation and optimisation of products in fragrance and food industries can be attributed to computer-aided design of experiment (DOE) methods / software, which allow for quicker and more economical product development.

The use of artificial intelligence (AI) and machine learning (ML) has further enhanced these techniques, with additional factors such as trendspotting, simulation of ingredient interactions, and use of biomimetic research being valued within the industry. AI-powered tools such as IMB and Symrise’s “Philyra”, or Givaudan’s “Carto” and focused research groups such as dsm-firmenich’s “d-lab” highlight that this multidisciplinary field is full of exciting prospects and potential for innovation.

ENVIRONMENTAL

Enviro-informatics focuses on applying computer science, data science and chemistry to answer environmental science questions. The use of informatic techniques such as data curation, collection, storage, manipulation and presentation aid the development of the environmental research field.

Enviro-informatics studies collect vast amounts of data, which can be analysed and visualised to allow for the most informed decision-making. Methods like remote sensing and statistical risk analysis are data-focused studies which benefit from the employment of cheminformatics. Modelling and simulation of chemical, biological and environmental systems also allow for the optimisation of processes, to benefit the environment.

From atomic level to functional material scale, enviro-informatics has aided areas of research such as pollution mitigation, waste management, greenhouse gases, ecosystem monitoring, water quality management, renewable clean energy resources, and forecasting natural disasters such as wildfire spread.

Artificial intelligence (AI) has been integrated into enviro-informatics research to aid sustainability. A report from Microsoft and PwCon “How AI can Enable a Sustainable Future” predicts that using AI for environmental applications can provide early warning signs of illegal deforestation saving up to 32 million hectares of forest globally by 2030, and also has the potential to reduce global greenhouse gas emission by around 1.5 - 4.0% by 2030, using satellite data and ground based sensors as data sources.

Whilst discussing the pro’s of using AI for environmental studies, it would be one-sided to not mention the environmental concerns that come along with it. Although considered a game-changer for the acceleration of innovation, the use of AI (including for cheminformatics studies) does give rise to concerns about its environmental impact. It is important to keep in mind the development, maintenance, and disposal of AI usage, and the carbon footprint that comes along with it. We understand that being responsible with our usage of AI is very important and think this is a great opportunity for the innovation of new and more energy efficient AI algorithms.

MATERIALS

Materials informatics combines data science, computer science and materials chemistry to analyse, understand and manipulate experimental and structural data.

Employing materials informatics to a workflow removes the need for the more costly and time consuming “trial and error” approach to material design. Materials informatics can be employed to analyse large datasets including materials descriptors (for example what atoms make up the molecule, or its electronegativity) and properties (such as hardness, density, conductivity, etc), focusing on data from accurate experiments to provide optimal solutions. Trained algorithms can suggest novel material candidates that can be screened for synthetic feasibility, sustainability, chemical intuition, and cost – again removing the need to use “trial and error” methods or optimise pre-established materials.

Materials informatics can be approached in a forward or inverse approach. Forwards approaches usually require the screening of different materials with different descriptors to find out what properties are gained. The inverse approach asks what properties are desired from a material and works backwards to design a material with descriptors that provide the required properties. Both methods of materials design using informatics relies on the “structure, property, processing, performance” relationships, which lies within trained algorithms.

The application of artificial intelligence (AI) and machine learning (ML) methods have aided the field by allowing for faster screening of vast databases, more efficient design of experiments (DoE), and ultimately the discovery of new materials. With larger datasets, with better quality of data and the use of AI / ML techniques, the informatics field has dramatically accelerated materials discovery.