| Cerius²·Ludi |
A Ludi de novo or link library consists of two files, a structure file and a targets file. For each fragment in the Ludi library, there is an entry in the structure file that defines the element types and the coordinates of the atoms in the fragment. The structure entry is a multi-record entry; that is, it consists of more than one line. The first line of the structure entry specifies the fragment label and the number of atom records in the entry. The fragment label uniquely identifies the fragment within the library. The remaining lines in the structure entry define the fragment topology. The format of the topology section may either follow the format of Brookhaven Protein Data Bank (PDB) atom records or the format of MDL Mol files (including the header block, counts line, atom block and bond block). Connectivity is deduced from the atomic coordinates. The number of atom records may exceed the number of atoms in the fragment as some atoms may be listed twice for reasons that are explained below.
During a Ludi run up to six fragment atoms or groups of fragment atoms (the target sites) are fit onto a model of the receptor (the interaction sites). For example, when a receptor active site contains a hydrogen bond donor group, acceptor interaction sites are generated during a Ludi run. These interaction sites define the acceptable positions for a hydrogen bond acceptor that is participating in the hydrogen bond. Fragments with hydrogen bond acceptor target sites are then fit onto these interaction sites.
For each fragment in the Ludi library there is an entry in the targets file that specifies the type of each target site in the fragment. There are two acceptable formats for the targets file: a single record format and a multi-record format. Table 11 describes the format of a single-record targets entry. The first value in the single-record targets entry is the fragment label. The second value is no longer used because it has been replaced by a user-specified parameter. The remaining values specify the chemical types of the fragment's target sites. Each fragment has a minimum of two, and a maximum of six, target sites. Some of the chemical types refer to two-atom target sites and some to one-atom target sites. For example, hydrogen bond donating groups are two-atom target sites. Hydrophobic atoms are one-atom target sites. A description of the chemical types, including their classification as one- or two-atom targets, is given in Table 12.
| Chemical Type Index | Number of Atoms |
Description |
Example |
|---|---|---|---|
| -3 | 1 | Hydrogen Bond Acceptor | Alcohol oxygen. |
| 1 | 1 | Hydrophobic Aliphatic | sp3 hybridized carbon. |
| 2 | 1 | Hydrophobic Aromatic | sp2 hybridized carbon. |
| 3 | 2 | Hydrogen Bond Acceptor | Carbonyl group. |
| 5 | 2 | Hydrogen Bond Donor | Polar hydrogen and heavy atom. |
| 7 | 2 | Link | Methyl group. |
|
Each entry in the single-record targets file corresponds to an entry in the structure file. Each value in the targets entry that specifies the chemical type of a target site corresponds to one or two records in the structure entry. The first chemical type value in the targets file corresponds to the first or the first two atom records in the structure entry. If the chemical type specifies a one-atom target, the correspondence is to the first atom record of the structure entry. If the chemical type specifies a two-atom target, the correspondence is to the first two atom records of the structure entry.
|
The last field of the first line of the structure file entry identifies the format of the library, either MOL_FORMAT or PDB_FORMAT. If the format tag is missing the library is assumed to be in PDB format.
|
Figure 6 describes the format for a multi-record targets file. The first line of the file is a header that specifies the version of Ludi_Genlib that produced the file. This line is followed by entries for each molecule in the library. The first record of the entry gives the library label and the number of atoms in the fragment. There is one line for each atom in the fragment that specifies atom number, atom name, accessible surface estimate, number of ludi types for the atom and the list of ludi types.
<run_name>.ludi_pdb, fort.10
This file describes the protein into whose active site Ludi is fitting fragments. The protein is specified by the Define ___ Model as Receptor parameter on the Ludi Receptor Based Design control panel. The file is in standard pdb format. This file may contain only the portion of the protein that falls with the search sphere specified at run time; it need not contain the entire protein.The link fort.10 points to the <run_name>.ludi_pdb when Ludi is run in standard mode (for the <run_name>.ludi_pdb file that is used in active analog mode, see the next section).
<run_name>.ludi_pdb, fort.3
This file describes the superimposed ligands for which Ludi is suggesting analogs. The superimposed ligands are defined by the Define Model as Active Analogs parameter on the Ludi Active Analogs Based Design control panel. The file is in pdb format.. Each ligand is separated from the previous ligand by a TER record. The link fort.3 points to the <run_name>.ludi_pdb file when Ludi is run in active analog mode.
<run_name>.ludi_inhib, fort. 7
This file describes the partially constructed molecule onto which Ludi tries to add fragments. The molecule is specified by the Define ___ Links from Selected Model parameters on the Ludi Design control panels. The <run_name>.ludi_inhib file is in standard pdb format. The link fort.7 points to the <run_name>.ludi_inhib file when Ludi is run in link mode. Link mode is specified by checking the Link Sites parameter on the Ludi Design control panels.
<run_name>.ludi_inp
When Ludi is run, an input file containing the Ludi execution parameters is automatically created. A typical example of the <run_name>.ludi_inp file follows:
TITLEEvery line (with the exception of the TITLE, IOUT and IBINRD lines) corresponds to a parameter in the Ludi Runtime Parameter control panel, the Ludi Design control panel or the Library Filters options.
CUTOFF 5.000000
XSITE -1.853000
YSITE 14.311000
ZSITE 16.658001
RMSMAX 0.300000
PRESEL 2.000000
VDWCUT 3.000000
ESCUT 2.500000
ANGMAX 0.000000
IOUT 0
IELEC 1
IDENSL 25
IDENSP 25
IFLAGV 1
ILINK 0
IANALG 0
IBIFUR 0
ICONMI 0
WLINK 1.000000
WLIPO 1.000000
WHBOND 1.000000
INEWSC 0
IMINSC 0
NHITS 999
ITARGT 0
IBINRD 0
IBURID 0
IGROUP 0
INOTGR 0
INOTAT 0
ICAVMX 0
INVERT 0
IROT 0
<run_name>.ludi_env
This file contains a set of Unix commands that set environment variables specifying the full pathnames of the library files. The environment variables are accessible only by the Ludi run background job and cease to exist when the Ludi run background job completes.
Output Files--Ludi run background job
fort.46
This file contains runtime status information. The contents of this file are printed in the textport when the Completion Status button on the Ludi Job Control control panel is clicked The following is typical of the status information contained in this file:
LUDI has found 7 molecules so far.
It has considered 17 ( 1%) molecules from the database.
The top score so far is 392.
<run_name>.ludi_pseudo_protein, fort.12
This file describes the exclusion shell that Ludi constructs from the active analogs. No fragment will be fit outside of this shell. The file is in PDB format and can be read into Cerius2 by using the Ludi Load control panel and checking the Analogs Shell button. The fort.12 file produced by the background job is automatically renamed to <run_name>.ludi_pseudo_protein when the background job completes.
<run_name>.ludi_scores, fort.47
This file contains a list of the hits from the Ludi run background job sorted by decreasing score. This file is only created if the background job runs to completion. There is one entry in the <run_name>.ludi_scores file for each hit. The fields of the <run_name>.ludi_scores file are described in Table 15. The following shows two typical entries:
file label score
-------------------------
69 G61 535 90 4
66 G35 437 106 3
| Column # | Heading | Description |
|---|---|---|
| 1 | file | File identifier for the hit.1 |
| 2 | label | Library identifier for the hit. |
| 3 | score | Score |
| 4 | Ratio of score to maximum possible score for the hit. | |
| 5 | Total volume of cavities formed by fit (cubic angstroms). |
There is one entry in the <run_name>.ludi_sum1 file for each hit. An entry may span several lines. The fields of the <run_name>.ludi_sum1 file are described in Table 16. The following is a typical entry:
hydrogen bonds
# label rmsd nfit natm contact occup R(O-H) <N-H.O
-----------------------------------------------------------
56 PYR 0.26 4 10 81
hydrophobic contacts: TRP 215 C
CYS 220 SG
hydrogen bonds : GLY 219 O 1.77 150.46
......................................................
<run_name>.ludi_sum2, fort.50
This file supplements and summarizes the information in <run_name>.ludi_sum1. The fort.50 file produced by the Ludi run background job is automatically renamed to <run_name>.ludi_sum2 when the background job completes.
# label rmsd nfit natm contact occup link lipoph. h-bonds sum hbond lipo link
-----------------------------------------------------------------------------
57 ASH 0.29 4 5 41 0 0 2 129 166 59 0
<run_name>.ludi_frag_<xxx>, fort.<yyy>
When the Ludi run background job finds a hit, it stores the hit in a file. The first hit is put in the file, fort.51, the next in fort.52, and so on. Subsequently the hits are culled so that only the highest scoring copy of a fragment is retained, therefore the numbering may not remain consecutive. When the Ludi run background job completes, the fort.<yyy> files are renamed by subtracting 50 from <yyy>, prefixing the result with one or two zeros to fill three places, and appending the result to <run_name>.ludi_frag. For example, file identifier 110 becomes <run_name>.ludi_frag060.
<run_name>.ludi_err, fort.49
Fort.49 contains any error messages produced by the Ludi run background job. It is automatically renamed to <run_name>.ludi_err.
<run_name>.ludi_log
This file contains information describing the run of the Ludi run background job. The fields are self-explanatory and include a list of the runtime parameters, numbers of interaction sites generated and a summary of fitting results for each fragment in the library.
<run_name>.ludi_targ_sites, fort.8
This file contains the list of target atoms for each fragment that was used in fitting to the interaction sites. If the library targets file format was of the single-record type, it will recapitulate what is already in the targets file. If the targets file format was of the multi-record format, this file lists for each fit: the fragment label and the atom name, atom number and type for each target atom.
AAAA N 2 -3 C 4 1 C 5 1
| Filenamea | Description |
|---|---|
| fort.10 | PDB file for the receptor. |
| fort.9 | PDB file for the ligand. |
| aThese files are deleted when the score has been computed. | |
<SCORE_RUN>.ludi_log contains information describing the run of the Ludi score background job. Fort.40 contains the scoring results and <SCORE_RUN>.lig_names contains the list of ligands being scored. At the completion of the Ludi score background job, the three files are compiled into <SCORE_RUN>.ludi_score and then they are deleted.