브라우저에서 단백질 언어 모델 임베딩을 탐색, 관계·유사도 누락·주석·라벨 전이 확인, 데이터 업로드 없음
Explore protein language model (pLM) embeddings in your browser — reveal relationships sequence similarity misses, overlay annotations, and transfer labels. Nothing uploaded.
데이터TypeScript · PythonGitHub @tsenoner온라인운영자 미확인
써 보셨어요? 한 번만 눌러 주세요. 로그인은 필요 없어요.
최근 완료된 8시간 회차 사이 GitHub 스타 증가량 순
Explore protein language model embeddings in your browser — surface relationships sequence similarity misses, overlay annotations, and transfer labels (EAT). Nothing uploaded.
README 전체를 불러오는 중입니다. 아래는 저장된 발췌입니다.
ProtSpace is a free, in-browser tool for exploring protein language model (pLM) embeddings. Project embeddings into an interactive map to surface relationships that sequence similarity misses, overlay biological annotations, transfer labels to unannotated proteins with a confidence score (EAT), inspect 3D structures, and export publication-ready figures, at Swiss-Prot scale, with your bundle never leaving the browser.README 전문 보기 ↗ProtSpace is a free, in-browser tool for exploring protein language model (pLM) embeddings. Project embeddings into an interactive map to surface relationships that sequence similarity misses, overlay biological annotations, transfer labels to unannotated proteins with a confidence score (EAT), inspect 3D structures, and export publication-ready figures, at Swiss-Prot scale, with your bundle never leaving the browser.
protspace.app, drag & drop a .parquetbundle file to start exploring. Loading, exploring and exporting a bundle run entirely client-side, the file never leaves your machine. You can also drop a .fasta for instant preparation on supported deployments; that path uploads your sequences to the ProtSpace prep backend, which computes the embeddings a browser cannot. See Is my data uploaded to a server? for details.
Option 1: Google Colab (no local installation)
Generate .parquetbundle files directly in your browser:
Option 2: ProtSpace Python package (local installation)
pip install protspace
# Build a visualization bundle from your embeddings (HDF5)
protspace prepare -i embeddings.h5 -m pca2,umap2 -o output
# ...or from a FASTA file (auto-embeds via the Biocentral API)
protspace prepare -i sequences.fasta -e prot_t5 -m pca2 -o output
See the ProtSpace Python package for the full CLI (annotation transfer, quality metrics, multiple pLMs, and more).
Full documentation, user guides, data preparation, and feature explanations.
If you use ProtSpace, please cite the web application preprint (latest):
Senoner, T., Vahidi, P., Olenyi, T., Senoner, F., Sisman, G., Kahl, E., Rost, B., & Koludarov, I. (2026). ProtSpace: Protein Universe in Your Browser. bioRxiv. https://doi.org/10.64898/2026.05.04.722720
The original, peer-reviewed ProtSpace publication:
Senoner, T., Olenyi, T., Heinzinger, M., Spannagl, A., Bouras, G., Rost, B., & Koludarov, I. (2025). ProtSpace: A Tool for Visualizing Protein Space. Journal of Molecular Biology, 437(15), 168940. https://doi.org/10.1016/j.jmb.2025.168940
A machine-readable CITATION.cff is included, use GitHub's "Cite this repository" button to export BibTeX or APA.