Functional Modules of openGPS
Predict
You can predict the protein localization and generate images, with the input of protein sequence and nuclear image (optional).
Submit
Scientists are welcomed to share and submit their own protein localization experiment images. If interested, you can compare them with the prediction images.
Introduction to deepGPS Framework
DeepGPS is a multimodal deep learning model which can predict protein subcellular localization by outputting both text labels and protein images simultaneously, with the input of protein sequences.
Architecture of deepGPS, a multimodal deep learning model