\n
![\"1\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F2875376%2Fbc35071c965a0e79d597858642a9d05c%2FScreen%20Shot%202024-04-26%20at%2010.04.44%20AM.png?generation=1714106122622656&alt=media\")
Update: Structure of building block_2 and building block_3 have been updated (Steps incorporated: Hydrogen addition and solvent molecule removal)
\nJust created a notebook to visualize different building blocks and the entire molecule structure in 3D. Hope it helps in getting a better perspective of how different building blocks and the entire structure look like.
\n\n1. Molecule_Smiles
\n
2. Building block_1
\n![\"Building](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F2875376%2F626bc2ad9786a15a2bb94b57348532ed%2FScreen%20Shot%202024-04-26%20at%2010.06.56%20AM.png?generation=1714106247540141&alt=media\")
3. Building block_2
\n![](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F2875376%2F02462b246387344d84ba0b062750f233%2FScreen%20Shot%202024-04-27%20at%209.00.00%20AM.png?generation=1714188655962137&alt=media\")
4. Building block_3
\n![](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F2875376%2Fb130c9a70b1d244b426f7b7c38301382%2FScreen%20Shot%202024-04-27%20at%209.01.27%20AM.png?generation=1714188725261017&alt=media\")
\nNote: This is for the first molecule in the train_parquet file.
Update: Structure of building block_2 and building block_3 have been updated (Steps incorporated: Hydrogen addition and solvent molecule removal)
\nJust created a notebook to visualize different building blocks and the entire molecule structure in 3D. Hope it helps in getting a better perspective of how different building blocks and the entire structure look like.
\n\n1. Molecule_Smiles
\n
2. Building block_1
\n
3. Building block_2
\n
4. Building block_3
\n
\nNote: This is for the first molecule in the train_parquet file.
Some of these structures (the last two) are not quite right.
\nPerhaps try Chem.AddHs(mol) before embedding a 3D coordinate system?
", "rawMarkdown": "Some of these structures (the last two) are not quite right.\n\nPerhaps try Chem.AddHs(mol) before embedding a 3D coordinate system?", "votes": null }, { "id": "2777893", "postDate": "04/26/2024 21:13:38", "content": "I'll bet that the last two have salts present - Cl (bright green in 2nd to last) and Br (dark red in last) - which should be removed before visualizing them. in both cases the Cl or Br atoms are not covalently bonded to the BB structure, but when the 3D structure was created, the ended up too close to other atoms. The visualizer assigns bonds by proximity and those atoms are too close to others, so they get bonds that aren't real.
", "rawMarkdown": "I'll bet that the last two have salts present - Cl (bright green in 2nd to last) and Br (dark red in last) - which should be removed before visualizing them. in both cases the Cl or Br atoms are not covalently bonded to the BB structure, but when the 3D structure was created, the ended up too close to other atoms. The visualizer assigns bonds by proximity and those atoms are too close to others, so they get bonds that aren't real.", "votes": null }, { "id": "2778122", "postDate": "04/27/2024 00:41:40", "content": "this is the binding process (docking)
\n(generated from https://github.com/luwei0917/DynamicBind)
![](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F113660%2F06d3aa78a0de2c9b2c7c79402ad57ad6%2FPeek%202024-04-27%2008-40.gif?generation=1714178464581473&alt=media\")
from this you can tell what feature are important (e.g. 3d shape, surface …)
", "rawMarkdown": "this is the binding process (docking)\n(generated from https://github.com/luwei0917/DynamicBind)\n\n\n\nfrom this you can tell what feature are important (e.g. 3d shape, surface ...)", "votes": null }, { "id": "2778182", "postDate": "04/27/2024 02:24:58", "content": "\nIt is a diffusion model approach. Interesting. DynamicBind
", "rawMarkdown": "~~Is that from docking or from Molecular Dynamics? ~~\n\nIt is a diffusion model approach. Interesting. [DynamicBind](https://www.nature.com/articles/s41467-024-45461-2)", "votes": null }, { "id": "2778240", "postDate": "04/27/2024 03:47:46", "content": "Hey! @chemdatafarmer @kirkdco
\nThank you so much for pointing it out. This is my first time working with SMILES, it was a great insight. I added H, removed Br and Cl and updated the structures. I would appreciate if you could once confirm that these are correct, else I will further rectify! (I lack expertise in cheminfo)
You're welcome, and those updated structures look very reasonable.
", "rawMarkdown": "You're welcome, and those updated structures look very reasonable.", "votes": null }, { "id": "2778252", "postDate": "04/27/2024 04:09:15", "content": "Thanks for confirming! :)
\nNotebook updated as well.
Some of these structures (the last two) are not quite right.
\nPerhaps try Chem.AddHs(mol) before embedding a 3D coordinate system?
", "votes": null, "replies": [ { "id": 2777893, "author_name": "kirkdco", "author_url": "", "post_date": "04/26/2024 21:13:38", "content": "I'll bet that the last two have salts present - Cl (bright green in 2nd to last) and Br (dark red in last) - which should be removed before visualizing them. in both cases the Cl or Br atoms are not covalently bonded to the BB structure, but when the 3D structure was created, the ended up too close to other atoms. The visualizer assigns bonds by proximity and those atoms are too close to others, so they get bonds that aren't real.
", "votes": null, "replies": [ { "id": 2778240, "author_name": "ahsuna123", "author_url": "", "post_date": "04/27/2024 03:47:46", "content": "Hey! @chemdatafarmer @kirkdco
\nThank you so much for pointing it out. This is my first time working with SMILES, it was a great insight. I added H, removed Br and Cl and updated the structures. I would appreciate if you could once confirm that these are correct, else I will further rectify! (I lack expertise in cheminfo)
You're welcome, and those updated structures look very reasonable.
", "votes": null, "replies": [ { "id": 2778252, "author_name": "ahsuna123", "author_url": "", "post_date": "04/27/2024 04:09:15", "content": "Thanks for confirming! :)
\nNotebook updated as well.
this is the binding process (docking)
\n(generated from https://github.com/luwei0917/DynamicBind)
from this you can tell what feature are important (e.g. 3d shape, surface …)
", "votes": null, "replies": [ { "id": 2778182, "author_name": "kirkdco", "author_url": "", "post_date": "04/27/2024 02:24:58", "content": "\nIt is a diffusion model approach. Interesting. DynamicBind
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "2776248": "**Update:** Structure of building block_2 and building block_3 have been updated (Steps incorporated: Hydrogen addition and solvent molecule removal)\n\nJust created a notebook to visualize different building blocks and the entire molecule structure in 3D. Hope it helps in getting a better perspective of how different building blocks and the entire structure look like.\n \n[Anusha's Research Book](https://www.kaggle.com/code/ahsuna123/anusha-s-research-book-3d-visualization) \n\n**1. Molecule_Smiles**\n\n\n**2. Building block_1**\n\n\n**3. Building block_2**\n\n\n**4. Building block_3**\n\n**Note:** This is for the first molecule in the train_parquet file.", "2776823": "Some of these structures (the last two) are not quite right.\n\nPerhaps try Chem.AddHs(mol) before embedding a 3D coordinate system?", "2777893": "I'll bet that the last two have salts present - Cl (bright green in 2nd to last) and Br (dark red in last) - which should be removed before visualizing them. in both cases the Cl or Br atoms are not covalently bonded to the BB structure, but when the 3D structure was created, the ended up too close to other atoms. The visualizer assigns bonds by proximity and those atoms are too close to others, so they get bonds that aren't real.", "2778122": "this is the binding process (docking)\n(generated from https://github.com/luwei0917/DynamicBind)\n\n\n\nfrom this you can tell what feature are important (e.g. 3d shape, surface ...)", "2778182": "~~Is that from docking or from Molecular Dynamics? ~~\n\nIt is a diffusion model approach. Interesting. [DynamicBind](https://www.nature.com/articles/s41467-024-45461-2)", "2778240": "Hey! @chemdatafarmer @kirkdco \nThank you so much for pointing it out. This is my first time working with SMILES, it was a great insight. I added H, removed Br and Cl and updated the structures. I would appreciate if you could once confirm that these are correct, else I will further rectify! (I lack expertise in cheminfo)", "2778244": "You're welcome, and those updated structures look very reasonable.", "2778252": "Thanks for confirming! :)\nNotebook updated as well." }, "source": "meta" }