tag:blogger.com,1999:blog-1136153439451224584.post9150552007860422911..comments2024-03-27T06:45:59.174-07:00Comments on Practical Fragments: Myriad metrics – but which are useful?Dr. Teddy Zhttp://www.blogger.com/profile/07288045760981372367noreply@blogger.comBlogger9125tag:blogger.com,1999:blog-1136153439451224584.post-65068761145715218842013-09-11T01:46:21.330-07:002013-09-11T01:46:21.330-07:00Anonymous got it right. In my view the Shultz pape...Anonymous got it right. In my view the Shultz paper is hot air as a result of not understanding why we use these metrics.<br /><br />We use these measures because us humans do not cope well with multi-objective optimisation. So we reduce space to two parameters 9potency & size, potency & lipophilicity). <br /><br />You use the measure that suits your problem - that is why I don't buy Shultz's argument that "LLE is best" for mathematical reasons. It won't work when I am trying to decrease the size of my mols, and likewise LE will not work when optimising lipophilicity (d'uh!).Willemnoreply@blogger.comtag:blogger.com,1999:blog-1136153439451224584.post-38315927307678462622013-08-30T16:07:04.675-07:002013-08-30T16:07:04.675-07:00Thanks for sharing such a great information.Am loo...Thanks for sharing such a great information.Am looking forward for your net post.Anonymoushttps://www.blogger.com/profile/14264532599014333308noreply@blogger.comtag:blogger.com,1999:blog-1136153439451224584.post-84336151738182330162013-08-26T14:01:45.227-07:002013-08-26T14:01:45.227-07:00The problem is not with the papers or the studies ...The problem is not with the papers or the studies (with the exception of the "Golden Ratio" paper). The problem is with implementation.<br /><br />All of these studies should aid the practicing medicinal chemists in how to think about things from multiple perspectives and focus on why structural changes result in activity differences, rather than "a methyl group made that number go up and the other number go down".<br /><br />Unfortunately, some (many) are written in a way that suggests that chemists should stop thinking and just follow the rules. <br /><br />Worth reading and thinking about, but not necessarily worth reading and acting upon.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-1136153439451224584.post-31626969540584748032013-08-26T11:46:35.249-07:002013-08-26T11:46:35.249-07:00I defined my LE, why should I use yours?I defined my LE, why should I use yours?Dr. Teddy Zhttps://www.blogger.com/profile/07288045760981372367noreply@blogger.comtag:blogger.com,1999:blog-1136153439451224584.post-17162988026657946142013-08-26T11:32:05.123-07:002013-08-26T11:32:05.123-07:00Dr. Teddy, try with LE = -RT*ln(IC50)/HAC ...Dr. Teddy, try with LE = -RT*ln(IC50)/HAC ...Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-1136153439451224584.post-6682496666755385982013-08-26T09:58:43.420-07:002013-08-26T09:58:43.420-07:00Thanks Anonymous, good catch - I've corrected ...Thanks Anonymous, good catch - I've corrected it above. However, the LE of Fragment 1 is still better, which would flag it as interesting despite the pathetic 1 millimolar IC50. LE can be useful for flagging such fragments that might otherwise be overlooked.Dan Erlansonhttps://www.blogger.com/profile/07927082337051189270noreply@blogger.comtag:blogger.com,1999:blog-1136153439451224584.post-2602062994793306452013-08-26T09:22:40.396-07:002013-08-26T09:22:40.396-07:00Using LE (pIC50/HAC) I get 3/8 =0.375
and 5/19 =0....Using LE (pIC50/HAC) I get 3/8 =0.375<br />and 5/19 =0.26.Dr. Teddy Zhttps://www.blogger.com/profile/07288045760981372367noreply@blogger.comtag:blogger.com,1999:blog-1136153439451224584.post-9607741673902575702013-08-26T04:58:09.010-07:002013-08-26T04:58:09.010-07:00For me, and I am sure I have said this before, LE ...For me, and I am sure I have said this before, LE is not meant to represent reality, which is what I think most metrics are trying to do. LE, OTOH, is a useful guide to help you decide if you are making smart, efficient use of chemistry space, rather than just glomming stuff on. I would never argue 0.29 is much worse than 0.31. I would say that 0.21 is much worse than 0.31 however.Dr. Teddy Zhttps://www.blogger.com/profile/07288045760981372367noreply@blogger.comtag:blogger.com,1999:blog-1136153439451224584.post-50968110536217807622013-08-26T00:03:03.209-07:002013-08-26T00:03:03.209-07:00Dan, based on your hypothetical data I calculate a...Dan, based on your hypothetical data I calculate an LE of 0.36 for Fragment 2, which would make it an OK starting point.Anonymousnoreply@blogger.com