We have collected some FAQs on LigBuilder V3 in the following section. Hope they will help you to use LigBuilder V3. If these FAQs still do not cover your question, please feel free to contact us at:
LigBuilder v1.0 first became available to the public in September, 1999. Since then, we have registered hundreds of users. As feedback, these users point out the bugs in the program and also make all kinds of comments and suggestion. All of these have been pushing us to go ahead.
Ten years past, we developed the new version: LigBuilder V2. Besides bug-fixing, major improvements include:
(1) A new design strategy "exploring mode" was introduced to extend the original growing and linking design strategy. With the exploring mode, the program will explore the drug chemical space automatically.
(2) We use multi seed structures to generate the initial generation rather than single seed structures. Furthermore, an ensemble growing and linking algorithm was applied to put a lot of seed into competition to selecting the rational seeds.
(3) Improved the score function.
(4) Developed a post recommender to refine design results.
(5) Developed an embedded synthesis analysis module to enhance the synthesis accessibility of design results.
(6) Much more user-friendly.
Another ten years past, we developed the new version: LigBuilder V3. Besides bug-fixing, major improvements include:
(1) Enable mutli-target design
(2) Learning from known inhibitor to design mimic compounds
(3) Better controling for realizing user's intention of design
(4) Improved the score function.
(5) Interface for load external programs in generating compounds (for example, using other scoring funcion, ADMET rule etc.)
(6) Prepared fragments database based on known drugs and bio-active compounds.
(7) Much more user-friendly.
Unfortunatly, the answer is "NO". Although DNA-based drug design shares the same concepts with protein-based drug design, there are two reasons preventing user from directly applying LigBuilder V3 to such a project: (1) New parameters are necessary for estimating the DNA-ligand interactions; (2) The binding site of a DNA (minor groove) is more open than a protein (usually a pocket), therefore the construction strategy may also need to be adjusted. We may develope this function in next version.
This is not so easy to interpret, user may have to inspect step by step. The exploring mode is much more robust with immoderate parameters, but the growing mode and linking mode is very susceptible. The problem may be caused by (1) The defination of binding site, so user should check the ID of binding site detected by Cavity module, and also check whether the selected binding site is correct by visually inspection. If Cavity could not generated desired binding site, user may force Cavity to generate the binding site directly(refer to the section of Cavity) (2) Unreasonable requirement of design, for example, the required pKd should be less than 8 in most cases. Especially, if the maximal binding affinity of binding site predicted by Cavity is low, user should not set a high value for target pKd. (3) For the growing mode or linking mode, improper seed structure cause most troubles. (a) user should check whether the seed structure is checmial reasonable. (b) whetehr the seed structure is in the binding site, and no atoms are out of the surface defined by Cavity. (c) user should check the growing site on the seed structure, or user could set all hydrogen to be converted to growing site in the parameter file(d) LigBuilder will provide diagnose information if the seed structure violate structure parameters of design, and user may modify the seed or change the parameters. Notice: some parameters are loaded by INCLUDE keyword, so user should define parameters by adding it to your own input file. We are strongly discourage user to modify any input file in the default directory. (4) the linking algorithm is more complicated than the growing algorithm. Linking between two separated structures will happen only when both conformation and chemistry are proper. Therefore, linking mode usually gives fewer results than growing mode does. Since the seed structures for linking may have no reasonable linkage between them can be made, because their positions or orientations are not "proper". So if this does happen, please modify your seed structure. Leave enough space between each fragment is a good idea. Besides, LigBuilder could use ensemble linking algorithm to select proper fragments for linking, so it is better to provide much more seed fragments to increase the possibility of success. (5) You could check the "active_TaskName.list" (e.g. active_1db4.list) to inspect the status of design, which may help you to find the reason. If you run LigBuilder in single CPU mode, you could check the status.lig file in result directory. (6) Please check whether the program is killed by system due to memory limitation. Using large fragment library or enable synthesis analysis will consume a lot of memory. If the program still gives no result because of some mysterious reasons, please let us know.
Maybe you have noticed that LigBuilder V3 gives estimation of the binding affinities of the resultant molecules. It does so by using the SCORE v4.0 algorithm. It is inherited from the SCORE series: Wang, R.; Liu, L.; Lai, L.; Tang, Y. "SCORE: A new empirical method for estimating the binding affinity of a protein-ligand complex", J.Mol.Model., 1998, 4, 379-394. If you want to estimate the binding affinity of a single ligand molecule to its receptor protein, you may use the "Score" function of Build.
LigBuilder V3 also estimates the octanol/water partition coefficients of the resultant molecules. It does so by using the XLOGP algorithm: Wang, R.; Gao, Y.; Lai, L. "Calculating partition coefficient by atom-additive method", Perspectives in Drug Design and Discovery, 2000, 19, 47-66. If you want to estimate the LogP value of a single molecule, you may download the XLOGP program from our website http://mdl.ipc.pku.edu.cn/cgi-bin/down.cgi
Good question!
LigBuilder V3 is a multiple-purposed structure-based drug design program. It provides the abilities for binding pocket analyzing, lead discovery, lead optimization, mimic design, de novo design, multi-target design, molecule filter and synthesis analysis. The modularized function of LigBuilder V3 enhances the flexibility in its application and the program provides a high degree of automatization to design molecules by one-stop process. Besides, it is very easy to use --- the users like this so much.
However, please remember: You'd better not use it as a black box although we try to streamline the design process; use it as an idea generator for instead. Especially when analyzing the final results given by LigBuilder V3, we do believe that expertise is still very important.
[Content] [Introduction] [Download] [Install] [Overview] [CAVITY] [BUILD] [Skills] [FAQs]
(These web pages are edited by Yaxia Yuan. Latest update: Jan., 2021)