wells-wood-research / timed-design

Protein Sequence Design with Deep Learning and Tooling like Monte Carlo Sampling and Analysis
46 stars 11 forks source link

Added help labels to the predicted sequence metrics. #58

Closed oliveramds closed 4 months ago

oliveramds commented 1 year ago

Still subject to change.

Charge:

Isoelectric Point:

Molecular Weight:

Mol. Ext. Coeff. @ 280 nm:

universvm commented 1 year ago

Hi @oliveramds ,

Have you looked at the functions from AMPAL that calculate these metrics? There may be details you can add (eg. pH).

I see that you define several variables. Have you tried running the UI and see if it works? Can you post screenshots?

oliveramds commented 1 year ago

Hi @universvm, I haven't yet looked at the functions on AMPAL, thank you for pointing it out. I'll take a look :)

Here are the screenshots using 1qys:

universvm commented 1 year ago

Thanks looks good! Maybe we need to specify that it is the designed protein vs the real one (wrt "this protein").

oliveramds commented 1 year ago

St.image does not support gifs. Had to import additional modules: st.components and base64 to open the gif file locally. Added gif_container as an empty st element, which is populated by the function _draw_gif with a div containing a gif, and is cleared upon starting the app after the button is clicked. Due to the gif’s size (27MB), it takes a few seconds to load. It also takes a few seconds to reload after the user selects a file because Streamlit has to re-render the whole page in its new appearance.

I would recommend using a video file because:

image
universvm commented 4 months ago

Closed as stale request. Will need to be updated before merging.