ProteinMPNN
ProteinMPNN · Papers & sources
Robust deep learning-based protein sequence design using ProteinMPNN
Wet-lab results
Ab002| Protein name | Target | Subset | Design workflow | SEC purity (%) | SDS purity (%) | Single-point BLI net response | Nonspecific binding signalBVP ELISA measures nonspecific antibody binding to baculovirus particles. The reported value is a background-corrected net signal. Under matching assay and normalization conditions, lower values generally indicate less nonspecific binding. The assay helps assess polyreactivity and the risk of rapid clearance in vivo.Assay reference |
|---|---|---|---|---|---|---|---|
| IgGM-IL7RA-009 | IL-7Rα | Ab002 | IgGM × ProteinMPNN | 97.8 | 98.6 | 0.008 | |
| IgGM-IL7RA-010 | IL-7Rα | Ab002 | IgGM × ProteinMPNN | 88.8 | 98.7 | -0.01 | |
| IgGM-IL7RA-011 | IL-7Rα | Ab002 | IgGM × ProteinMPNN | 80.4 | 90.8 | -0.0082 | |
| IgGM-IL7RA-012 | IL-7Rα | Ab002 | IgGM × ProteinMPNN | 95 | 95.5 | -0.0015 | |
| BoltzGen-IL7RA-021 | IL-7Rα | Ab002 | BoltzGen × ProteinMPNN | 95.3 | 98.2 | -0.0008 | |
| BoltzGen-IL7RA-022 | IL-7Rα | Ab002 | BoltzGen × ProteinMPNN | 96.2 | 95.1 | 0.0939 | |
| BoltzGen-IL7RA-023 | IL-7Rα | Ab002 | BoltzGen × ProteinMPNN | Aggregation | 90.2 | 0.0014 | |
| BoltzGen-IL7RA-024 | IL-7Rα | Ab002 | BoltzGen × ProteinMPNN | Aggregation | 29.6 | 0.0279 | |
| BoltzGen-IL7RA-025 | IL-7Rα | Ab002 | BoltzGen × ProteinMPNN | 95.4 | 98.8 | 0.2371 | -0.066395 |
| BoltzGen-IL7RA-026 | IL-7Rα | Ab002 | BoltzGen × ProteinMPNN | 98.5 | 94.9 | 0 | |
| BoltzGen-IL7RA-027 | IL-7Rα | Ab002 | BoltzGen × ProteinMPNN | Aggregation | 98.3 | 0.0071 | |
| BoltzGen-IL7RA-028 | IL-7Rα | Ab002 | BoltzGen × ProteinMPNN | 91.2 | 95.9 | -0.0037 | |
| BoltzGen-IL7RA-029 | IL-7Rα | Ab002 | BoltzGen × ProteinMPNN | 90.1 | 98.3 | 0.1791 | -0.085895 |
| BoltzGen-IL7RA-030 | IL-7Rα | Ab002 | BoltzGen × ProteinMPNN | 77.6 | 95.8 | 0.0249 | |
| BoltzGen-IL7RA-031 | IL-7Rα | Ab002 | BoltzGen × ProteinMPNN | Aggregation | 0.0043 | ||
| BoltzGen-IL7RA-032 | IL-7Rα | Ab002 | BoltzGen × ProteinMPNN | 93.3 | 98.6 | -0.0032 | |
| BoltzGen-IL7RA-033 | IL-7Rα | Ab002 | BoltzGen × ProteinMPNN | 93.6 | 97.5 | 0.0022 | |
| BoltzGen-IL7RA-034 | IL-7Rα | Ab002 | BoltzGen × ProteinMPNN | 94.4 | 99 | 0.0103 | |
| BoltzGen-IL7RA-035 | IL-7Rα | Ab002 | BoltzGen × ProteinMPNN | 87.1 | 95.9 | 0.2547 | |
| BoltzGen-IL7RA-036 | IL-7Rα | Ab002 | BoltzGen × ProteinMPNN | 51 | 71.6 18.8 | 0.0508 | |
| BoltzGen-IL7RA-037 | IL-7Rα | Ab002 | BoltzGen × ProteinMPNN | 95.9 | 96.8 | 0.0101 | |
| BoltzGen-IL7RA-038 | IL-7Rα | Ab002 | BoltzGen × ProteinMPNN | 64.5 | 31.3 67.1 | 0.0252 | |
| BoltzGen-IL7RA-039 | IL-7Rα | Ab002 | BoltzGen × ProteinMPNN | Aggregation | 0.0033 | ||
| BoltzGen-IL7RA-040 | IL-7Rα | Ab002 | BoltzGen × ProteinMPNN | 90.2 | 97.9 | 0.1922 | -0.034995 |
| RFantibody-IL7RA-021 | IL-7Rα | Ab002 | RFantibody × ProteinMPNN | 95.2 | 96.1 | -0.0001 |
1–25 of 64 candidates
Compute resources and costs
RESOURCES / MODEL STAGE
IL7RA · Model-level aggregateModel-stage resources
Device with recorded maximum:NVIDIA RTX PRO 5000 72GB Blackwell
Devices in sampled records:NVIDIA GeForce RTX 4090 / NVIDIA RTX PRO 5000 72GB Blackwell
Memory values are device sampling maxima, not continuous monitoring maxima. GPU pricing classes do not verify hardware; unknown hardware is priced as RTX 4090. Costs exclude 30 GPU tasks shorter than 600 s (1.0558 GPU·h); CPU-only records remain included and queue time is excluded.
IL‑7Rα project cost breakdown
IL7RA / PROJECT BUDGET
¥13,847.43 CNYEstimated project cost breakdown
Totals cover all models and evaluation stages, independently of the selected model. CPU core-hours and GPU hours are priced separately.
Structure generation
¥1,497.2210.8%
Sequence design
¥213.381.5%
Rosetta
¥5,096.936.8%
AF2
¥5,415.9439.1%
AF3
¥1,623.9911.7%