Phenix LigandFit RSCC calculation questions
Dear Phenix developers, I hope this message finds you well. We had two specific questions about how the real-space correlation coefficient (RSCC) is computed when using Phenix.LigandFit: 1. What radius is used around the ligand or search center when calculating the RSCC during ligand fitting? Is this radius fixed or adaptive? 2. Which CCTBX function(s) underlie the RSCC calculation used in LigandFit? We aim to compare the placements from LigandFit to other ligand placement methods, and your response would be appreciated. Best regards, -- *Theodore W. Teng* Harvard College | Class of 2027 A.B. student in Chemical and Physical Biology Secondary in Global Health and Health Policy
There is a modified quote from a famous statistician, which one might keep
in mind on such ventures:
All ligands are wrong, but some are useful.
Cheers, BR
------------------------------------------------------
Bernhard Rupp
https://www.hofkristallamt.org/
[email protected]
[email protected]
+1 925 209 7429
+43 676 571 0536
------------------------------------------------------
People can be divided into three classes:
The few who make things happen;
The many who watch what is happening;
and the overwhelming majority who have
no clue what is happening.
------------------------------------------------------
On Tue, Jul 29, 2025 at 11:35 AM Theo Teng
Dear Phenix developers,
I hope this message finds you well.
We had two specific questions about how the real-space correlation coefficient (RSCC) is computed when using Phenix.LigandFit:
1. What radius is used around the ligand or search center when calculating the RSCC during ligand fitting? Is this radius fixed or adaptive?
2. Which CCTBX function(s) underlie the RSCC calculation used in LigandFit?
We aim to compare the placements from LigandFit to other ligand placement methods, and your response would be appreciated.
Best regards, --
*Theodore W. Teng* Harvard College | Class of 2027 A.B. student in Chemical and Physical Biology Secondary in Global Health and Health Policy _______________________________________________ phenixbb mailing list -- [email protected] To unsubscribe send an email to [email protected] Unsubscribe: phenixbb-leave@%(host_name)s
Hi Theo, quick answer: you need CCmask for your purpose[*] and I don't know exactly what version of CC LigandFit computes and reports. Longer answer: for this kind of benchmarking, it is essential to use the exact same definition of CC and the exact same code (with identical settings) to compute it. If you are comfortable with Python scripting and familiar with CCTBX, I can send you a script that demonstrates how to do this. Pavel [*] https://journals.iucr.org/d/issues/2018/09/00/kw5139/index.html On 7/29/25 11:34, Theo Teng wrote:
Dear Phenix developers,
I hope this message finds you well.
We had two specific questions about how the real-space correlation coefficient (RSCC) is computed when using Phenix.LigandFit:
1. What radius is used around the ligand or search center when calculating the RSCC during ligand fitting? Is this radius fixed or adaptive?
2. Which CCTBX function(s) underlie the RSCC calculation used in LigandFit?
We aim to compare the placements from LigandFit to other ligand placement methods, and your response would be appreciated.
Best regards,
participants (3)
-
Bernhard Rupp -
Pavel Afonine -
Theo Teng