Applications of AI in the pharma sector

Biostate AI was founded in 2023 and is based in California and Texas. The start-up works with academic, nonprofit, and industry partners to develop and implement generative AI uses that benefit human health.  

In July 2024, Biostate AI released its Total RNA Sequencing technology suite, able to analyze all types of RNA, including non-coding RNA species, which all play important roles in biological processes. This is different from traditional RNA sequencing, which mainly captures information from mRNA. The Total RNA Sequencing technology enables extensive RNA data collection via their patent-pending Barcode-Integrated Reverse Transcription (BIRT) technology. This technology aims to better understand gene expression and regulation, improving comprehension of disease mechanisms and potential therapeutic targets.  

In 2024, Biostate AI further launched the OmicsWeb Copilot, which is a conversational AI tool used to help biologists analyze and visualize data efficiently. The OmicsWeb Copilot currently uses advanced LLMs to understand requests and make personalized software and scripts for data analysis.  

Furthermore, Biostate AI hopes this AI Copilot platform will help predict human and animal health changes, including toxicity and drug efficacy responses. Indeed, it demonstrated in 2024 that RNA expression taken from rats’ blood before giving them a drug could accurately predict their survival. To size up this proof of concept to be able to predict the toxicity of drugs in humans, a lot more data must be collected, analyzed, and fed into the AI models for training. Biostate’s tools are part of a bigger trend of using AI in biosciences. The tools can address challenges in collecting and analyzing biological data. AI could potentially reduce the need for animal testing and manage data complexity.  

As of 2024, Biostate AI has raised $4 million in funding from investors such as Catapult VC and Vision Plus Capital. It has also collaborated with Caltech and Twist Biosciences to further the predictive capabilities of AI in health science. Biostate AI also plans to work with hospital biorepositories, pharma and biotech companies, and academic researchers to scale their multi-omic (constituents within a cell) data collection and AI training.

Current gene editing approaches can face challenges such as brute-force mutagenesis, resulting in low success rates. Berkeley-based start-up Profluent has used GenAI technologies to develop a gene-editing system named OpenCRISPR-1. The start-up uses LLMs fed with extensive biological data to enhance existing gene-editing techniques, specifically the well-known clustered regularly interspaced short palindromic repeats (CRISPR) technology.  

Profluent trained its AI gene editor by inputting massive-scale DNA sequences and biological context to make millions of varied CRISPR-like proteins that do not appear in nature, thereby increasing the diversity of all known CRISPR families. These AI-customized gene-editing proteins are proteins inspired by natural structures but with unique molecular compositions.  

They have exhibited enhanced precision and efficiency in editing the human genome, reducing off-target effects and genetic inconsistencies. With more than 400 mutations compared to traditional CRISPR tools, OpenCRISPR-1 demonstrates a median rate of unwanted genetic modifications of less than 1%, making it a viable alternative for experiments requiring precise gene editing capabilities.  

This shift from accidental discovery to intentional design enables precise and controlled genetic modifications, potentially revolutionizing therapeutic development for complex societal challenges. Additionally, Profluent has raised a total of $44 million, with Spark Capital leading a $35 million financing round in March 2024. The funding will support Profluent's growth, particularly in advancing its proprietary LLMs and datasets, improving wet lab capabilities, and enhancing CRISPR gene editing. This investment aims to enable the development and validation of innovative proteins, potentially advancing healthcare and disease treatment.