Products and AI
BioRecommender boasts an extensive database of over 7 million products, a significant portion of which are associated with official gene symbols, as recognized by the HUGO Gene Nomenclature Committee (HGNC).
For instance, when searching for "TNF alpha," you must use the correct nomenclature such as "TNF" or "TNF beta" to access the full array of recognized names and aliases for a protein.
For instance, when searching for "TNF alpha," you must use the correct nomenclature such as "TNF" or "TNF beta" to access the full array of recognized names and aliases for a protein.
Upon selecting a gene, based on a customer's name, two distinct series emerge:
1. Customer-Cited Genes/Proteins: These are genes or proteins the customer has mentioned in their scientific publications. Applying a date filter will also refine this gene list accordingly.
2. Recommender Genes/Proteins: Derived from the GeneRecommender Engine, this series utilizes a proprietary AI-driven neural network to suggest genes or proteins closely related to those already identified.
Predicting Genes with AI:
The GeneRecommender's neural network is trained on millions of papers, from which it extracts genes, diseases, and significant keywords. It calculates the "distance" between genes in a 64-dimensional space, performing billions of calculations to score genes based on their proximity to each other. This method mirrors the approach researchers use to determine their next focus of study, offering a probability of correctly identifying a researcher's gene of interest between 30-50%.
Selecting a recommended gene allows users to choose the producing company, view the product SKU, and directly visit the product's webpage for more information. This functionality is invaluable both for promoting distributed or produced products and for understanding competitive offerings.
Generecommender
GeneRecommender can be integrated into your operations in two main ways:
1. Sponsorship: By sponsoring the GeneRecommender website (which is free for non-profit research), your banners and products will be featured on the platform, directing customers to your website for the recommended genes/products.
2. API Integration: Incorporating the GeneRecommender API into your ERP/CRM systems allows for the translation of customer-requested SKUs into genes. The engine then recommends corresponding genes/SKUs from your product range, potentially increasing orders by up to 30%.
Collaborative Efforts and Target Audiences:
TheProphetAI collaborates with esteemed institutions like Politecnico di Milano and Università di Milano Bicocca to refine its AI capabilities, targeting a wide range of clients from antibody, protein, and ELISA kit producers to life science researchers and pharmaceutical companies with R&D divisions. This diverse approach ensures that TheProphetAI's innovative solutions reach a broad spectrum of the scientific and commercial community, driving forward the integration of AI in life sciences.
Main Targets (Customers)
1) Antibody, protein, kit ELISA producers:
This target segment comprises 50 major companies and hundreds of medium and small-sized enterprises. The estimated total market size, focusing solely on antigens (antibodies, proteins, and ELISA kits), is substantial.
2) Distributors:
These distributors play a critical role in the supply chain, connecting target 1 companies with the broader market. TheProphetAI's offerings can also benefit these entities.
3) Life-Science Researchers and Pharmaceutical Companies with R&D divisions:
In essence, we provide an AI system built on proprietary neural networks capable of suggesting a series of genes (proteins) when given an initial list of other genes (proteins). These protein recommendations are generated through the neural network's analysis of millions of scientific papers, representing a collective knowledge repository.
These results are accessible through a user-friendly web interface (www.generecommender.com) or via APIs that facilitate the creation of gene lists with corresponding scores.
These scores can be translated into company product codes, used to generate PDF reports for customer distribution, and much more.
These scores can be translated into company product codes, used to generate PDF reports for customer distribution, and much more.
All companies can leverage our APIs to automate the bundling of new products based on customer requests. They can draw from their historical customer data or integrate our solutions into their CRM/ERP systems.
Technical Development of TheProphetAi and capabilities
Main skill:
1) LLM technology thanks to University of Milan Bicocca Collaboration, one year project with the help of the grant of Regione Lombardia Fast Tech.
2) Proprietary neural network able to recommend the best protein is possible
3) NLP competence
MIND development of a NLP pipeline to extract gene, disease and other biological entities citations from scientific texts (articles or abstracts) using state-of-the-art transformer-based models
DCB Optimize DP2 model accuracy on less cited genes through data balancing techniques and Identification and implementation of an efficient hyperparameter optimization framework for the DP2 deep learning model
Grant and Awards
TheProphetAi won a grant from Regione Lombardia (ID 3156798, 701)