Ai-driven gene expression evaluation with protoplast assays

Overview

Phytoform Labs offers an advanced solution for evaluating gene expression through a combination of AI-driven genomic modeling and high-throughput protoplast assays. This innovative approach allows for scalable evaluation of regulatory DNA sequences, offering both in silico predictions and in vivo validations. By leveraging a sequence-to-function genomic model trained on extensive multi-species genomic data, alongside a protoplast-based massively parallel reporter assay (MPRA), Phytoform enables efficient discovery of novel plant promoter variants.

Technical specifications

The core technology involves a sequence-to-function (S2F) genomic model trained on 12 species reference genomes and over 5,000 RNA-seq coverage tracks. This model predicts gene expression profiles from 65kb genomic sequences. Complementarily, the protoplast-based MPRA quantitatively validates thousands of gene sequences simultaneously, associating RNA-barcodes with DNA regulatory variants using the MPRAflow pipeline. The model's effectiveness is demonstrated by a high gene level correlation and conditional correlation, validated through custom assays and fluorescence-based transient expression assays.

Technology readiness level

Currently, the technology is at TRL 4, having been validated under laboratory conditions with both computational and experimental methods. The next steps involve promoter evolution campaigns and adjusting project plans to focus on specific needs, such as screening existing promoter variants.