![](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F15928403%2F094c9409481a12ddd7c0462fa2110e61%2FScreenshot%202024-04-23%20at%203.22.45PM.png?generation=1713866017013915&alt=media\")
I am very new to biochemistry and trying to clear out few basic doubts that I have. I am confused on the Triazine (scaffold) part and DNA Encoded Library (DEL).
\nTo begin with, the DEL part as described by Brenner, Lerner 1992 paper, the following is how DEL is constructed:
\n
Now this part is clear to me, that once a binding protein is detected, we read the attached DNA sequence, and find the identity of the protein.
\nNow coming to Triazine scaffold part, here is one drawing that I can get:
\n![](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F15928403%2F26111e21ad714970448de554a8043bdf%2FScreenshot%202024-04-23%20at%203.19.38PM.png?generation=1713866328001705&alt=media\")
If I compare this structure with the structure by Brenner, I see some issues. First of all, there are now two peptide chains (say R1 = N(R3)^2 and R2=N(R4)^2) emanating from the Triazine core. But we only have one DNA sequence.
\n1) So how would I identify two peptide sequences from single DNA sequence?
\n2) Moroever, what is the purpose of Triazine core?
\n3) Also say one peptide chain (R1) out of two interacts with target protein, but the presence of a second adjacent protein chain (R2) - wont that create steric clash between say R1 protein and target protein?? Unless there are two binding pockets in the target protein, and binding to two pockets is essential for activation/inhibition.
\n4) In the training data, there is full molecule and building blocks, why not use the full molecule structure,. why we need the building blocks structure as well??
Apologies if my question seem too mundane!
", "messages": [ { "id": "2769384", "postDate": "04/23/2024 10:08:53", "content": "I am very new to biochemistry and trying to clear out few basic doubts that I have. I am confused on the Triazine (scaffold) part and DNA Encoded Library (DEL).
\nTo begin with, the DEL part as described by Brenner, Lerner 1992 paper, the following is how DEL is constructed:
\n
Now this part is clear to me, that once a binding protein is detected, we read the attached DNA sequence, and find the identity of the protein.
\nNow coming to Triazine scaffold part, here is one drawing that I can get:
\n
If I compare this structure with the structure by Brenner, I see some issues. First of all, there are now two peptide chains (say R1 = N(R3)^2 and R2=N(R4)^2) emanating from the Triazine core. But we only have one DNA sequence.
\n1) So how would I identify two peptide sequences from single DNA sequence?
\n2) Moroever, what is the purpose of Triazine core?
\n3) Also say one peptide chain (R1) out of two interacts with target protein, but the presence of a second adjacent protein chain (R2) - wont that create steric clash between say R1 protein and target protein?? Unless there are two binding pockets in the target protein, and binding to two pockets is essential for activation/inhibition.
\n4) In the training data, there is full molecule and building blocks, why not use the full molecule structure,. why we need the building blocks structure as well??
Apologies if my question seem too mundane!
", "rawMarkdown": "I am very new to biochemistry and trying to clear out few basic doubts that I have. I am confused on the Triazine (scaffold) part and DNA Encoded Library (DEL). \n\nTo begin with, the DEL part as described by Brenner, Lerner 1992 paper, the following is how DEL is constructed:\n\n\n\nNow this part is clear to me, that once a binding protein is detected, we read the attached DNA sequence, and find the identity of the protein.\n\nNow coming to Triazine scaffold part, here is one drawing that I can get:\n\n\n\nIf I compare this structure with the structure by Brenner, I see some issues. First of all, there are now two peptide chains (say R1 = N(R3)^2 and R2=N(R4)^2) emanating from the Triazine core. But we only have one DNA sequence. \n\n1) So how would I identify two peptide sequences from single DNA sequence? \n2) Moroever, what is the purpose of Triazine core? \n3) Also say one peptide chain (R1) out of two interacts with target protein, but the presence of a second adjacent protein chain (R2) - wont that create steric clash between say R1 protein and target protein?? Unless there are two binding pockets in the target protein, and binding to two pockets is essential for activation/inhibition.\n4) In the training data, there is full molecule and building blocks, why not use the full molecule structure,. why we need the building blocks structure as well??\n\nApologies if my question seem too mundane!", "votes": null }, { "id": "2769529", "postDate": "04/23/2024 11:46:41", "content": "1) You won't, you identify the whole Triazine molecule.
\nEDIT: Second read, I assume you do each of the three proteins separately. So if you find the DNA codebar you assign a unique protein to the linked molecule (triazine in that case).
\n2) Is just the result of use such building blocks (BB).
\n3) I think the model has to deal with any binding mechanism. And has to predict it by the data we have.
\n4) Is just more specific. You are free to neglect any information that you conseider redundant for your approach.
The 'R's are not polipeptid chains, or proteins, they are just sidechains of the molecule.
", "rawMarkdown": "The 'R's are not polipeptid chains, or proteins, they are just sidechains of the molecule.", "votes": null } ], "comments": [ { "id": 2769529, "author_name": "sacuscreed", "author_url": "", "post_date": "04/23/2024 11:46:41", "content": "1) You won't, you identify the whole Triazine molecule.
\nEDIT: Second read, I assume you do each of the three proteins separately. So if you find the DNA codebar you assign a unique protein to the linked molecule (triazine in that case).
\n2) Is just the result of use such building blocks (BB).
\n3) I think the model has to deal with any binding mechanism. And has to predict it by the data we have.
\n4) Is just more specific. You are free to neglect any information that you conseider redundant for your approach.
The 'R's are not polipeptid chains, or proteins, they are just sidechains of the molecule.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "2769384": "I am very new to biochemistry and trying to clear out few basic doubts that I have. I am confused on the Triazine (scaffold) part and DNA Encoded Library (DEL). \n\nTo begin with, the DEL part as described by Brenner, Lerner 1992 paper, the following is how DEL is constructed:\n\n\n\nNow this part is clear to me, that once a binding protein is detected, we read the attached DNA sequence, and find the identity of the protein.\n\nNow coming to Triazine scaffold part, here is one drawing that I can get:\n\n\n\nIf I compare this structure with the structure by Brenner, I see some issues. First of all, there are now two peptide chains (say R1 = N(R3)^2 and R2=N(R4)^2) emanating from the Triazine core. But we only have one DNA sequence. \n\n1) So how would I identify two peptide sequences from single DNA sequence? \n2) Moroever, what is the purpose of Triazine core? \n3) Also say one peptide chain (R1) out of two interacts with target protein, but the presence of a second adjacent protein chain (R2) - wont that create steric clash between say R1 protein and target protein?? Unless there are two binding pockets in the target protein, and binding to two pockets is essential for activation/inhibition.\n4) In the training data, there is full molecule and building blocks, why not use the full molecule structure,. why we need the building blocks structure as well??\n\nApologies if my question seem too mundane!", "2769529": "1) You won't, you identify the whole Triazine molecule.\nEDIT: Second read, I assume you do each of the three proteins separately. So if you find the DNA codebar you assign a unique protein to the linked molecule (triazine in that case).\n2) Is just the result of use such building blocks (BB).\n3) I think the model has to deal with any binding mechanism. And has to predict it by the data we have.\n4) Is just more specific. You are free to neglect any information that you conseider redundant for your approach.", "2771732": "The 'R's are not polipeptid chains, or proteins, they are just sidechains of the molecule." }, "source": "meta" }