Part A (From Pranam)

Sign up for HuggingFace (we will be using PepMLM: https://huggingface.co/ChatterjeeLab/PepMLM-650M)

  1. Once you login, go to the page (https://huggingface.co/settings/tokens). Click +Create new token.
  2. Make sure you type the full name ChatterjeeLab/PepMLM-650M when searching for repos. Click save token and you will see the newly token (copy that).
  3. Go to the page (https://huggingface.co/ChatterjeeLab/PepMLM-650M) and find their Colab Notebook (link).
  4. Make a copy to your Google Drive, choose T4 GPU and run each block.
  5. When running into the block Input HF token , a pop-up will show Enter your token (input will not be visible):. Paste your token and Add token as git credential? (Y/n) choose n.

Yessss. Done all that.

Find the amino acid sequence for SOD1 in UniProt (ID: P00441), a protein when mutated, can cause Amyotrophic lateral sclerosis (ALS). In fact, the A4V (when you change position 4 from Alanine to Valine) causes the most aggressive form of ALS, so make that change in the sequence

Uhu, alanine is not in number 4 position but number 5.

1.png

Mutated the alanine of position 5 to valine.

Enter your mutated SOD1 sequence into the PepMLM inference API and generate 4 peptides of length 12 amino acids (Step 5 takes a while so you can also just pick 1 or 2 peptides)

2.png

3.png

To your list, add this known SOD1-binding peptide to your list: FLYRWLPSRRGG [from -https://genesdev.cshlp.org/content/22/11/1451]

Couldn’t access genesdev.

4.png

Original SOD1 and mutated SOD1, the binder peptides generated (binder 0-3) and given above (binder 4).

5.png

To make the count easier, I number them as: