Log In|Sign Up
ENIntroduction
AAV ViaPredictor is a collaborative development between Cyagen Biosciences and the AI Innovation Center of Tsinghua Panyu Research Institute. It is an AI-based AAV yield prediction tool that integrates multi-dimensional feature data with deep learning frameworks. This tool enables accurate and efficient yield optimization support for gene therapy drug research and production by predicting yields across different serotype vectors.
Operating Instructions
The tool accepts base sequences for AAV2 serotype to perform yield predictions. Specific application conditions are as follows:
Figure 1. Operation diagram, where the input fragment can only be nucleotide sequence.
AAV2 Serotype
Mutation support range
Only supports inserting 21-30 base pairs (7-10 amino acids) between amino acid positions 587-588 of the wild-type sequence.
Reference sequence
Wild-type sequence (positions 561-592): DEEEIRTTNPVATEQYGSVSTNLQRGN^RQAAT (the "^" symbol marks the insertion site between positions 587-588).
Insertion structure: Automatically appends an N-terminal cleavage sequence "TG" and a C-terminal cleavage sequence "GLS" to form "TG~GLS" format, where "~" represents the input segment.
Inputs must be nucleotide sequences, not amino acid sequences.
Result Interpretation
The result will be presented as a data table. The "sequence" column shows the input sequences, and the "score" column displays the model-predicted yield scores (The score has been scaled using log2). Higher scores indicate higher production yields, as illustrated below:
| sequence | score |
|---|---|
| AAV2 WT | 1.13 |
| input sequences | prediction score |
