Case Studies

Proven Results

Explore our AI-designed therapeutics with interactive 3D molecular viewers. Rotate, zoom, and examine the structures that are advancing medicine.

Click and drag on molecules to rotate • Scroll to zoom

Antiviral Discovery

COVID-19 Drug Discovery: CMP-Neu5Ac

University Lab

Identified Cytidine-5'-monophospho-N-acetylneuraminic acid sodium salt (CMP-Neu5Ac) as a SARS-CoV-2 Spike Protein inhibitor through AI-driven molecular docking. Wet lab validated by University Materials Science Lab.

SMILES Structure

CC(=O)N[C@@H]1[C@@H](O)C[C@@](OC2O[C@H](COP(O)(=O)O)[C@@H](O)[C@H]2O)(C(O)=O)O[C@H]1[C@H](O)[C@H](O)CO

Validated

Wet Lab Confirmation

Spike RBD

Target Binding Site

In Vitro

Confirmed Inhibition

Loading 3D structure...

Interactive 3D

Drug Repurposing

COVID-19 Drug Repurposing: FDA-Approved Antivirals

AttentionSite-Augmented DeepDTA

Top FDA-approved antiviral drugs predicted by AttentionSite-Augmented DeepDTA model to have highest affinity scores with 4 SARS-CoV-2 genome sequences. Identified candidates like Fosamprenavir, Elvitegravir, and Ritonavir as potential COVID-19 treatments.

3416

FDA Drugs Screened

Viral Targets

Antivirals Ranked

Drug Name	pIC50	FDA Rank	Antiviral Rank
1 Oseltamivir phosphate	8.29	#42	#1
2 Etravirine	8.22	#52	#2
3 Fosamprenavir calcium	8.20	#57	#3
4 Tipranavir	7.94	#108	#4
5 Elvitegravir	7.88	#141	#5

Top 5 candidates for Spike-ACE2 target

AI Predictions

De Novo Peptide Design

GPR-139 Selective Agonist: De Novo Design

DeepBio Scientific AI Platform

De novo designed GPR-139 selective agonist with 40% better predicted binding score than Dynorphin. Unlike RFDiffusion (21% structural similarity), our design achieves 0% similarity to existing proteins, enabling optimization to avoid off-target binding to opioid receptors (OPRM, OPRK, OPRD).

Peptide Sequence

VAFVKNWLEFLAQWVASRLVNHQIGLQQELQIVAQAVLRYHQAETAKLSYTLAN

+40%

Better Binding vs Dynorphin

Structural Similarity

Zero

Off-Target Opioid Binding

Loading 3D structure...

Interactive 3D

mRNA Language Model

Equi-mRNA: mRNA Expression & Stability Prediction

NeurIPS 2025 Publication

Equi-mRNA is the first codon-level equivariant mRNA language model that explicitly encodes synonymous codon symmetries using SO(2) group theory. With only 15M parameters, it outperforms models 100x larger across multiple mRNA benchmarks.

0.855

Best mRFP Score

0.764

Best Tc-Ribo Score

15M

Parameters (vs 1.6B)

Benchmark Results

Model	E.coli	MLOS	iCodon	Tc-Ribo	mRFP
Nucleotide-Based
RNA-FM	0.43	-	0.34	0.58	0.80
Aido mRNA	0.576	0.504	0.472	0.492	0.683
mRNA-FM	-	0.509	0.458	0.690	0.564
Codon-Based
CodonBert	0.57	0.543	0.350	0.502	0.832
HELM (MLM)	-	0.701	0.525	0.626	0.822
Equi-mRNA (5M)	0.581	0.705	0.519	0.764	0.853
Equi-mRNA (15M)	0.613	0.710	0.537	0.737	0.855

Bold = Best performance on dataset

NeurIPS 2025 Paper

Ready to Start Your Success Story?

Join leading pharmaceutical and biotech companies using DeepBio Scientific to accelerate therapeutic discovery with AI-powered molecular design.

Start a Partnership