DreamFold develops generative machine learning algorithms, including generative flow networks, to design and profile novel proteins for drug discovery. Their technology reduces the risk of late-stage failures in drug development by utilizing a feedback loop from biological assays to inform molecular selection for clinical trials.
Funding
$6M raised to dateRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
CDMIOOPVFounders
Product
Problem
The traditional drug discovery process is costly and time-consuming, often resulting in late-stage failures due to unforeseen issues with efficacy, toxicity, or patient heterogeneity. Identifying promising drug candidates early in the development pipeline remains a significant challenge.
Solution
DreamFold utilizes generative machine learning algorithms, including generative flow networks, to design and profile novel proteins for drug discovery. By integrating feedback from biological assays into the molecular design process, DreamFold's technology aims to reduce the risk of late-stage failures. The platform combines machine learning algorithms with scalable biological assays to learn patterns in the data. This approach accounts for tumor and patient heterogeneity, enabling scientists to make informed decisions and select the appropriate patient groups for clinical trials based on molecular data.
Target Audience
DreamFold's primary customers are pharmaceutical companies and research institutions involved in drug discovery and development.
Features
- Generative flow networks for de novo protein design
- Conditional flow matching algorithms for SE(3) equivariant motions
- Integration of synthetic biology data to inform model training
- Scalable biological assays for profiling millions of novel proteins
- Machine learning models that incorporate feedback from binding assays, functional cell assays, and single-cell patient-derived samples
- Patient stratification based on molecular data, including binding and microfluidics
- Cloud-based computation across 1000 GPUs, leveraging AnyScale, Nvidia, AWS, and Google