Replies: 9 comments 6 replies
-
|
Hi @LierB 1884: Schotten found that he could prepare N-benzoyl piperidine from piperidine with NaOH. My reference is Barbara Czako's et al. book "Named Reactions in Organic Synthesis" So yes, there are two main components to these reactions that have different functional groups and would be worth to separate. This list is proving to be difficult to complete so I propose we maybe switch out strategy.
For example to understand what I mean: atom_classes = {
'1': 'hydrogen',
'2': 'alkyl',
'3': 'amide',
'4: 'alcohol',
'5': 'aryl',
'6': 'benzyl',
'7': 'amine',
}We can perhaps use it like so in the SMILES string like so: The 4 and 7 meaning it can be alcohol or amine. What are your thoughts? Makes it easier to write. |
Beta Was this translation helpful? Give feedback.
-
|
Good morning @Sulstice
how can we be sure it takes |
Beta Was this translation helpful? Give feedback.
-
|
Hey @Sulstice first thanks for history class ;) I agree with you that switching strategy could be beneficial for this list. Including R groups is a good idea and with your example it gets very clear what you mean. We would need a whole list of all functional groups, which I could prepare (taking care of the order is not necessary, right). I have assumed now, that we also include R groups in reactions where different products are not of relevance (just to shorten down the string?). Or do you think this would be an overkill and we should stick to the way we do it now for "normal" cases? With this, what comes to my mind is, that for Schotten-Baumann the product depends on the reactant (2 reactions with two different reactants), but e.g. for Cannizzaro there are two products (with similar yields) in one reaction, which was not considered yet and could also be included. Ok maybe I'm anyways thinking too far away from the actual idea, please let me know about the points I mentioned. |
Beta Was this translation helpful? Give feedback.
-
|
hey @LierB I think this would be beneficial, as some reactions take place using specific amines (secondary, tertiary, primary) to yield different products. This is also the case to primary carbon, secondary carbon, tertiary carbon Also (for the reactions that yield 2 different products) I feel it would be beneficial to note the major product as well as the minor product, and 50:50 racemization if that is the case. hope this makes sense! |
Beta Was this translation helpful? Give feedback.
-
|
Hey @Safari222 yeah this was my thought to include them specifically, because in some cases actually just one/not all of them is/are possible to react. The question is now when we want to keep the Of course this makes sense! For now I just wrote the major product, before @Sulstice had the idea with the R groups. But I also agree that sometimes both products should be mentioned. Same for reactants in some cases, e.g. the Biginelli reaction where all 3 reactants are relevant. |
Beta Was this translation helpful? Give feedback.
-
|
@Safari222 Aziza, yeah it is. Go back to basics Sn2, Sn1, E1, E2. Lets use the atom class idea in conjunction with an actual indexing that makes the most sense and serves the purpose of filling out this list. For multicomponents we need to keep track a little bit. In the class object we need to have the attribute: So we can keep track of those. The reactants for something like Biginelli:
We can do In the smiles. This is saying that these 3 reactions to the product. For multicomponent reactions is there always only one dominant product is that a rule? |
Beta Was this translation helpful? Give feedback.
-
|
replying to @Sulstice 's question: "For multicomponent reactions is there always only one dominant product is that a rule?" In the literature it's emphasized that more than 2 products (usually 3 to 7) are converted to 1 product (or a single product, or the final product) or that there are almost no side products. It's a one-pot reaction and almost all atoms are incorporated in the product. They are seen as efficient reactions belonging to green chemistry. here is the DOI https://doi.org/10.1016/B978-0-12-817825-6.00002-1 to following book chapter intro here is a review introduction I read more and they all agreed on that, so I would say it's a rule. |
Beta Was this translation helpful? Give feedback.
-
|
Yeah Doemling comes from Sharpless and Ugi. He's reported up to 10-component reactions once as I think a theory before in this paper 3-10 is a good number:
https://www.sciencedirect.com/science/article/abs/pii/S1367593100000958?via%3Dihub Here's the figure. The rule for an MCR would be to be considered is: 1.) There is one dominant product meaning in the code the last molecule is going to the product regardless of how many components. 2.) For multi functional groups that have similar mechanisms based on the reaction. This might or not work but we can have atom classes virtually bonded Another alternative is to start using CXSMILES which is used for CGenFF atom typing to store metadata about the atom on the SMILES string. An example would be for perfluorobutanoic acid: Your guys thoughts? |
Beta Was this translation helpful? Give feedback.
-
|
ok so even up to 10 reactants, wow! 1.) or am I wrong? Because there it would be like 2.) with I think it gets more and more complex, but eventually it will include everything we need. |
Beta Was this translation helpful? Give feedback.
-
|
Yeah, so @LierB let's be simple for now to get majority of the list done. We will have to truncate reactions that have more than one product while I decide algorithmically how to unravel that information but still retain a 1-D format. Any reactions that are chemical_list = {
'reaction_1': 'R1.R2.P1',
'reaction_2': 'R1.P1',
'reaction_3': 'R1.R2.R3.R4.P1',
}Any other reactions that have multiple dominant products I can handle them. CXSMILES is useful for storing meta data on an atom type but the string itself is ugly and pollutes the output. Have to see. Try completing as much as you can with only 1 product and let's see how many we get. |
Beta Was this translation helpful? Give feedback.
-
|
Ok right @Sulstice , I will try to get the list done as soon as I can;) I will stick to one product for now (like we started, just with adding more reactants if relevant). In cases where there are actually more products relevant, I will decide for one as before and will additionally include the second as a comment. Later, if we want to switch to another approach it will be easier to change, if we have the info there I think. Do you agree? |
Beta Was this translation helpful? Give feedback.
-
|
Yeah, I think that's the best course of action. |
Beta Was this translation helpful? Give feedback.
-
|
@LierB Thank you for production of the list. Give me a little second and let's do a mass analysis on the overall picture together. We will be using a software called MolPDF. I think we can use that to measure performance as a visual inspection for list generation. The PR is maintained here: https://github.com/Sulstice/global-chem/pull/108 |
Beta Was this translation helpful? Give feedback.
-
|
So this is interesting and worth noting. Pericyclic reactions were a thing of the past for relic drug hunters but they still stand as valuable ring systems because of their spherical shape that is deep enough to fit into the pocket of a protein. For example: Adamantane:
When I saw the Wurtz Coupling it reminds me that we should look for Valuable Pericylic Reactions. Because this is just a cool application of that method.
We would need someone in the field that knows these reactions to make a comment. Think I know someone. |
Beta Was this translation helpful? Give feedback.
-
|
Okay, let's have some fun. It is interesting to me that to drug designers who consider themselves good can distinguish between valuable molecules. Adamantane is on my list as a common fragment and in my arsenal, it also exists in others and is demonstrated by CGenFF. CGenFF can show the adamantane as something valuable for drug design would be accounted for. So when I test it, it is fine. But pericylic reactions especially wurtz coupling has died out of value. Especially something as strained cyclobutane. The effective parameter worth looking at: Could be relatively and easy target. Name: CGenFF: Name: Adamantane CGenFF: |
Beta Was this translation helpful? Give feedback.
-
I have started with adding smiles to named_reactions_in_organic_synthesis.
For the Schotten-Baumann reaction and the Curtius rearrangement (up to now) there are more products possible dependent on the reactants. Schotten-Baumann will yield amides from amines or esters from alcohols. For now I have decided for one, but I think they would both be relevant I think (for Curtius even three products).
What would you say, should we add more products with different reactants like 'schotten_baumann_reaction_amines' and 'schotten_baumann_reaction_alcohols'? Or would someone has a different idea how to deal with such cases?
Grateful for replies including suggestions and any other issues and thoughts regarding the named reactions.
Betty
Beta Was this translation helpful? Give feedback.
All reactions