COSMOfrag Manual
Contents
1. COSMOlogic Predicting Solutions COSMOfrag Release 3 5 User s Manual by COSMOlogic GmbH amp Co KG A Klamt Imbacher Weg 46 D 51379 Leverkusen Germany Phone 49 2171 731 681 Fax 49 2171 731 689 E mail info cosmologic de Web _shttp www cosmologic de 1 4 ui 6 Contents NERO ELUTEE TE a ce cacti E ecceccvssivecavgevcdacussits eivsstectevsdhes save A ETE EEO O N E 1 1 1 CFDB the COSMOfrag Database ccccccccccecessesessececeeecesseseaeeeeeeecesseseaaeaeeeescessessaaeaeeeesens 2 1 2 COSMOsim Extension to Bioisoster Screening ssssssssssssesrsssrrrsrrrrrrrrrrrrrrrrrrrrrrrererrrererees 3 MaE RA ST E E E E E E E T E O E E EA E 3 RUNNING COSMOFrag ied icc siriene teneinde nereh ea aaaea dacs an a Aai cuiatac easdecduanintadisks lavevbuidlactate saces 4 3 1 Program exeCUtiONu atso acetate eck ee aa aa aa E a aa aaa aa eaaa i a aa 4 3 2 NAELA E i EE E E E E E A E E E 4 3 2 1 Global command NES easa aaan aeaa Aaa AATE KEES AARETE KEA AERAR TA KAEA EEEE 4 3 2 2 Structure tile aT eaa AAE KESAR oe aAA UKAA PEASE REEE a a KAARO TRAAT AAA VANA TEAS 9 3 3 GOSMO therm calc la Hons aeien onana andeka NARREN ies sseieacvactsaadeuueiseesucusdievaaterd EEREN 9 3 3 1 COSMOth rm iNPUt SECTION ccccccccccecessesssaecececesessnsaeseeeeeceseeseaaeseeeescessesaaaeseeeessesensnaaeess 9 3 3 2 Property prediction by COSMOtHESIIM ccccccccccecsssessnnecececsscssesusaee2. Human Serum Albumin e Octanol Water Partition Coefficient logPOW For solubilities or activity coefficients the SCREEN flag should be set in the COSMOtherm template input section not documented in the COSMOtherm manual which speeds up the calculation of these properties The 1 octanol water partition coefficient logkKOW is of special interest for the drug design process Based on the statistical analysis of several validation studies the prediction of the logkOW value is improved by the command logp_amine_corr 0 6 1 1 1 5 The NMETA n n 1 9 option can be used to generate several fragmentations of a compound Each of the fragmentations is stored into a separate metafile and the property calculation is done for each metafile of the compound Note that if another fragmentation of equivalent quality in terms of atom wise similarity as indicated by the MAXSTRING is not available the number of fragmentations written to metafiles can be smaller than the number indicated by the NMETA keyword In the results file filename res the results of the property calculation are tabulated for each fragmentation together with the average value and the standard deviation of the calculated values 11 3 4 Multiple jobs Multiple COSMOfrag jobs may be combined within a single input file The input for each successive job is separated from that of the preceding job step by a line of the form S Note that an empty row terminates proces
3. 1 6 functionality molecules are added virtually to the database taking set keywords into account MINSADD MIN MAXNUMHEV Virtually added molecules are taken into consideration for the database representation of molecules following on the list no cosmo files necessary Add really to CFDB covers ACTION 1 6 functionality molecules are added to the database taking set keywords into account MINSADD MIN MAXNUMHEV cosmo files have to be provided and will be stored as compressed cosmo file ccf in the database given with the CFDBDIR keyword Remove molecules from database covers ACTION 1 6 functionality o database molecules corresponding to the molecules on the list will be removed instead of structure files a list of unique names can be given CFDBDIR STRDIR LICENSEDIR ADDSIM ACCEPT_CHARGES ALLOW_ALKALI ALLOW_ZW ALLOW_RARE CHECKN X CCFCOM CFDB_LIST nnn CFCOSMO named CT_EXT CTCALC named relative or absolute path to the CFDB relative or absolute path to the structure files directory path to license file license ctd If not specified COSMOfrag searches in CFDBDIR add to CFDB also similar molecules stereo isomers allows for charges in the structure ACTION 1 functionality only allows alkali atoms in the compounds ACTION 1 functionality Restricted to W_EZSMI and USMI only No Warnings and Errors are written to the tab fie switches ZWITTERION detection to warning allows rar
4. COSMOsim Bioisosteric Similarity Based on COSMO RS F Profiles Michael Thormann Andreas Klamt Martin Hornig and Michael Almstetter J Chem Inf Model 46 1040 1053 2006 COSMO RS From Quantum Chemistry to Fluid Phase Thermodynamics and Drug Design A Klamt Elsevier Amsterdam 2005 8 Troubleshooting and Support The present COSMOfrag Version is subject to permanent methodical development and software optimization We explicitly request for sending us bug reports and feature suggestions Contact Prof Dr Andreas Klamt COSMOlogic GmbH amp Co KG Imbacher Weg 46 51379 Leverkusen Germany klamt cosmologic de and or info cosmologic de 49 0 2171 731681 18
5. almost any kind of calculation which can be done with COSMO files in COSMOtherm can be done similarly using COSMOfrag meta files Additionally COSMOfrag can be used to generate new COSMOfrag databases add molecules to an existing database or compress COSMO files by almost a factor of 20 in order to allow for an efficient storage of the CFDB 1 2 COSMOsim Extension to Bioisoster Screening The success of the COSMO RS theory has proven that o profiles hold the crucial information for most ADME properties as solubility blood brain partition coefficients intestinal absorption and even for many adsorption phenomena Considering the fundamental importance of the o profiles for surface interactions of molecules in liquid states they obviously also carry a large part of information required for the estimation of desolvation and binding processes which are responsible for the inhibition of enzyme receptors by drug molecules Thus a high similarity with respect to the o profiles appears to be an important condition for drugs of similar physiological activity Driven by this insight we have developed a o profile based drug similarity measure denoted o match similatity or SMS for the detection of new bioisosteric drug candidates The program extension COSMOsim enables the efficient calculation of this similarity for large libraries making use of the COSMOfrag technology In several examples COSMOsim has demonstrated its statistical and pharmaceutic
6. file the numbers of double bonds with possible cis trans isomerism and numbers of stereo atoms In combination with WRTDESCRIPT specifies the maximum number of meta files which are generated per compound instead of 1 Values can range from 1 to 9 Only fragmentations of equivalent quality to the best possible fragmentation are written into meta files Thus it is possible that less meta files per compound are written than indicated by the keyword indicates not to write identity meta files where possible but to use an alternative fragmentation leave one out principle POLYMER STREZCHIR TRY_CHARGE USESMINAM USENAM lt identifier gt USEOLDNAM USMILES USMILES_CHIR USMILES_EZ W_EZSMI WRTDESCRIPT WCAR WXYZ WSDF WCCF WMCOSCAR marks a polymer repeat unit by lodine cap atoms If lodine is used otherwise you may specify another halogen by F Cl or Br writes a string to the out file which specifies for 3D compounds R S and E Z for stero centres and double bonds respectively F indicates a failed stero analysis tries total charges of 1 if no neutral structure can be assigned use the name after the SMILES code as molecular name tag in multiple SD files to use for naming when splitting up single structure files metafiles indicates that the original COSMO files shall be referenced in meta files not the ccf files from CFDB generate a unique SMILES string for each file
7. gt Filesystem operations gt 1 file open error gt 2 file read error gt 3 file format not known gt 4 name assignment failed gt 5 file open error occured gt 6 filename or SMILES too long max 299 letters 12 gt Errors in SMILES code gt gt gt 16 20 31 SMILES contains more than one molecule Error reading SMILES string unresolved atom in SMILES gt Structures possibly valid but presently not supported by COSMOfrag gt VVVVV VV VV V gt 10 12 17 18 19 21 22 30 32 39 40 41 number of rings exceeds maximum more than 50 intramolecular donor acceptor pair found element unknown or not allowed residual charges in molecule molecule charged molecule too large parts of the molecule not connected isotope specified biradical specified no kekule structure found after ion loop compound most likely will be zwitterionic in water The keyword ALLOW_ZW triggers on the calculation of zwitterions molecule has a large ring which cannot be well handled by COSOMfrag The keyword LARGERING 999 activates the calculation of larger rings gt CEDB offers no acceptable fragment for atoms parts of the molecule no image for one atom aromaticity of some rings could not be defined invalid hybridization no atoms found only hydrogens in molecule single aromatic bond error in hydrogen addition invalid valence bond order trace back failed traceback failed in ring a
8. that the nth compound line will be read from the temp tab file output of the total COSMO charge This is useful for the evaluation of the fragmentation quality Writes some useful additional information to the tab file information on special geometrical characteristics is written to the out file possible internal H bonds lone pair conflicts planar NH2 trans O COX_group reads molecular charge from ccf file default when using CFDB_LIST sets the limit for the larger ring detection Cyclic compounds with a ring size below the threshold are treated as a normal compound in COSMOfrag If the threshold is exceeded in case ACTION 6 7 calculations the compound is omitted due to the large ring error error code 41 In case of ACTION 1 in combination with WRTDESCRIPT a string LARGERING is written to the tab file The default is LARGERING 13 Set nnn 999 to deactivate specifies the maximum number of non hydrogen atoms in a molecule to be accepted for CFDB addition output a string with maximum similarities of each atom in order of the atom numbering indicates that the 12 letter unique name shall be used for meta file labelling automatically set for SMILES structures specifies the minimum number of non hydrogen atoms in a molecule to be accepted for CFDB addition default is 1 specifies the threshold value of the minimum similarity in a molecule for CFDB addition default is 2 Values can range from 1 to 7 writes to the tab
9. COSMOsim parameters can be altered but we do not recommend doing so By supplying a value of a lt 1 d0 the usage of the Tanimoto index as suggested by Thormann can be triggered The keywords wprf named causes the program to write the o profiles of all compounds in the structure file list to jobname prf without calculating the SMS If wprf named is used the metafiles generated 16 during fragmentation are named with the compound name instead of temporary metafiles i e this option produces one mcos file per compound in the structure file list NOPRF No jobname prf file is written only the sorted list jobname prf2 is written Useful if the o profile database already exists All formats supported by COSMOfrag can be used in combination with COSMOsim Usually COSMOfrag will generate a fragmentation of the compound i e a mcos file and will generate a o profile from this fragmentation If the file format is a COSMO file or a compressed COSMO file ccf then the o profile is generated directly from this file without fragmentation In a COSMOsim run files with the extension prf or prf2 can also be within the compound list In that case the o profile required for the calculation of the SMS coefficient is not newly generated but read from the prf or prf2 files which is much faster Thus it is possible to re use previously generated o profiles from one or more COSMOsim jobs in later similarity searches This will be extremely f
10. Linux systems are case sensitive ACTION e 1 e 6 e 7 e 8 e 9 e 10 On OX e EE o Oi O Oy 0i Ob Oi O O This keyword determines the mode in which COSMOfrag is run Operations on listed molecules without database access very fast Parsing of the molecule and unique name determination File format conversions possible WCAR WXYZ Various checks of the validity of the geometry possible GEOCHECK Determination of descriptors possible WRTDESCRIPT Splitting of multiple SMILES SD files with USENAM Generation of SMILES strings for all possible EZ isomers possible W_EZSMI Operations on charged compounds and compounds with alkali atom are possible ALLOW_ALKALI and ACCEPT_CHARGE instead of structure files a list of unique names can be given Search database for molecules covers ACTION 1 functionality representation of the molecule within the database is checked identity minimum maximum similarity listed in statistics ouput printout of MAXSTRING possible instead of structure files a list of unique names can be given Fragmentation run covers ACTION 1 6 functionality COSMOtherm calculations possible ctcalc COSMOsim calculations possible COSMOsim generates metafiles mcos usable keywords NO_META_IDENT META_UNIQUENAME for named metafiles for the structure file list the named option can be added to any keyword Add virtually to database for CFDB enlargement screenings covers ACTION
11. RTDESCRIPT can be used for that task the number of atoms is available then from the table file The effect would be that smaller molecules are processed first and a higher probability for smaller molecules being selected for COSMO calculations results When the COSMO calculations are done the COSMO files should be added to the CFDB ACTION 9 6 COSMOsim The calculation of compound similarities based on the similarity of COSMO o profiles is triggered in the COSMOfrag input file by the keyword COSMOsim or COSMOsim ntarget nbest Here ntarget defines the number of compounds to be used as target By default only the first compound in the structure file list will be considered as target However it may be useful to search for the maximum average similarity of compounds to a set of ntarget target compounds which must be listed at the head of the structure file list For each compound in the list the average similarity as well as the individual similarities will be written to an output list j ooname prf There is no limitation on the number of compounds to be processed in one job nbest is the number of most similar compounds which will be written to a sorted list jobname prf2 The maximum possible value for nbest is 10000 By default the similarity is calculated as the sigma match similarity index SMS with the default parameters as described in the COSMOsim publication With the keyword SMS a b c d The
12. al plausibility its practicability for real drug research projects and its unique independence from the chemical structure The latter can be regarded as scaffold hopping in a natural way COSMOsim is integrated in the COSMOfrag program and it can be activated via a special license 2 Installation The COSMOfrag distribution contains the executable program the COSMOfrag database CFDB and some examples We recommend setting up such a file directory tree binLinux cf Linux executable binWindows cf exe Windows executable software COSMOfrag binMac cf MacOS executable COSMOfrag data base including CFDB subdirectories Extract the COSMOfrag database CFDB zip to the folder software COSMOfrag All subdirectories are automatically created To install COSMOfrag on a Linux Mac or Windows system copy the appropriate executable to a local directory and set the path If using Bourne Again Shell bash on a Linux system add to your bashrc file export PATH SPATH software COSMOfrag binLinux In a Windows environment set the path from the system control From the Windows Start menue select Control Panel then System Select the Advanced tab With the Environment Variables button at the bottom of the window you get a dialogue where you can edit several variables Edit the Path variable from the lower part and add the path to the directory where the COSMOfrag executable cf exe is located If
13. ast because for the similarity search only the similarity coefficient has to be calculated A database of precalculated o profiles of more than 8 8 million public available compounds is available from COSMOlogic If you have access to this database we recommend installing it to the folder software COSMOfrag CS_screeningDB and using the NOPRF keyword 7 References COSMOfrag A Novel Tool for High Throughput ADME Property Prediction and Similarity Screening Based on Quantum Chemistry Martin Hornig Andreas Klamt J Chem Inf Model 45 1169 1177 2005 Prediction of aqueous solubility of drugs and pesticides with COSMO RS Andreas Klamt Frank Eckert Martin Hornig Michael E Beck and Thorsten Burger J Comp Chem 23 275 281 2002 COSMO RS a novel view to physiological solvation and partition questions Andreas Klamt Frank Eckert and Martin Hornig Journal of Computer Aided Molecular Design 15 355 365 2001 Fast Solvent Screening via Quantum Chemistry COSMO RS Approach Frank Eckert and Andreas Klamt AICHE Journal 48 369 385 2002 17 Prediction of Blood Brain Partitioning and Human Serum Albumin Binding Based on COSMO RS o Moments Karin Wichmann Michael Diedenhofen and Andreas Klamt J Chem Inf Model 47 228 233 2007 Use of Surface Charges from DFT Calculation to Predict Intestinal Absorption Ron Jones Paul C Connolly Andreas Klamt and Michael Diedenhofen J Chem Inf Model 45 1337 1342 2005
14. dual molecules it allows for physically most sound estimations of general vapour liquid and liquid liquid equilibria and of related properties like solubilities and partition coefficients In addition it has been extended to properties like drug and pesticide solubility blood brain partition coefficients intestinal absorption soil sorption coefficients etc which are of importance in the design and development of drugs pesticides and other physiological agents Since thermodynamic calculations only require fractions of a second the overall timing of a COSMO RS investigation is mainly determined by the time demand of the underlying quantum chemical calculations for the molecules In the field of the design of drugs pesticides and other agents shortly called drug design below often a large variety of hundreds to several ten thousands of potential drug candidates has to be pre screened regarding their physiochemical properties each of them being typically in the range of a molecular weight MW of 300 500 i e having about 25 40 non hydrogen atoms A pre screening for large numbers of compounds using quantum chemical methods is almost unfeasible even on large parallel computer clusters In order to make COSMO RS applicable to high throughput calculations COSMOfrag has been developed The basic idea of COSMOfrag is to avoid the time consuming calculation of the screening charge densities o profiles by DFT COSMO calculations for each individua
15. e elements in the compounds The rare elements available are Sn Sb Te Hg Bi checks for C N and N N doublebonds which are not in a ring system and prints the configuration into the out file read a comment for the header of ccf files from line 4 of the input and places it in the ccf header calculation of the first nnn compounds of the CFDB the complete CFDB is calculated by CFDB_LIST Keyword has to be in the structure file list triggers writing of temp_cf cosmo Use CPCOSMO NAMED for getting lt name gt _cf cosmo run an external COSMOtherm program do a COSMOtherm calculation for each compound in the structure file list The fourth line of the input must be Sstart CT_input and the template input for the COSMOtherm calculation is closed by a line Send CT_input The input output and and table files of the COSMOtherm calculation are named temp inp temp out and temp tab and the metafile is named temp mcos If the CTCALC named option is used for each structure file e g molecule sdf the COSMOtherm calculation files will be named with the name of the structure file molecule inp molecule out and molecule tab and the metafile will be named molecule mcos Note that this option will create four files per file in your structure file list CTTABCOMP n CTtotCH FULLTAB GEOCHECK GET_CCFCHARGE LARGERING nnn MAXNUMHEV MAXSTRING META_UNIQUENAME MINNUMHEV MINSADD NEZCHIR NMETA NO_META_IDENT triggers
16. e memory at this time but no COSMO file is needed Consecutively the molecule just added will be taken into account for the representation of the other molecules on the list So if one molecule of a structurally similar group is added virtually the rest of the group can be represented by a fragment of this very molecule e To perform an operation like this another flag has to be set in the input file If MINSADD 1 is set in a virtual run only those molecules are virtually added to the CFDB whose smallest similarity of an atom is lt 1 Thus the first molecule of a group of molecules with a certain structural feature which is not yet represented in the CFDB is added For the following molecules of the group the minimum similarity is gt 1 for this special atom or group due to the molecule added before Then the MINSADD flag prevents those molecules from being virtually added as well 15 In summary a run on a complete set of molecules with ACTION 8 and MINSADD 1 should be done virtual runs are very fast Afterwards all compounds virtually added to the CFDB have to be selected there is a flag in the table file for added and calculated quantum chemically One thing can be done additionally to reduce computer time In case there are compounds in the set which are rather large i e more than 40 50 non hydrogen atoms the list of molecules can be ordered from smaller to larger ones before the virtual run COSMOfrag with ACTION 1 and W
17. ececeseeseaeseeeeseuesessaaees 11 3 4 Multiple jobs siscsc teats SA alls Sita as BB Bie as Mees 12 3 5 COSMO frag Koin 01i i TEE E E E E E 12 35 17 Metafile ga e a sides ae aaar a a aaa E aE ao On aaa ae aSa A nah ET PETS 12 3 5 2 Errors ANd Warnings cccccccccessssssssecececessessaeceeececesseseeaeseeeessessesaeaeeeesessseseaseeeeeeseeesesaaeess 12 The COSMOfrag database ccccccccssssssccccecessesssesecececsssesseaeeeeeesseeseasaeeeeeceseesaaaeaeseesessseseaeaeeeseess 14 APPliCatiOns choses ETE E os diigs dd dense a soda E E otihgs snes teesds haba divvidheedibassbiesstaeke 15 5 1 Visualization of metafiles 2a cs ssecs ss cdascedesadlecsuevanesccsasecadsesshedecssiccessatanesecasitcdduesshedeeanioess 15 5 2 Expansion of the COSMOfrag Database CFDB cc cccccsssccceesssceceessececeeseceeeesaeeeseesaeeeeees 15 COSMO SIM aaeeea Reeth be es aI ages ene Tae Gd anata tua ene aa Toes 16 REPEREN COS E A E RE NE AEE AE TA E E EET ET A E 17 Troubleshooting and Support ccccsscccccecessessnsesecececessesseaesececsseeseeaeeesecesseseaaeaeeeesessseseaeaeeeesens 18 1 Introduction COSMOfrag a fast shortcut for high throughput COSMO RS calculations The COSMO RS method has become an efficient and versatile tool for the prediction of a large variety of physicochemical properties especially in its efficient implementation within the COSMOtherm program Based on quantum chemical DFT COSMO calculations for the indivi
18. ere have identical hash codes up to the n order Indeed even some more distant information may be included in the n sphere since information is preferentially propagated along conjugated bonds assuming similarity along such bonds to 1 be more important than similarity along single bonds The highest order similarity code taken into account presently is 7 Including the 0 order we thus have 8 coefficients per atom Some additional less local information is gathered in 2 by coefficients per atom resulting in a total number of 10 coefficients per atom 1 1 CFDB the COSMOfrag Database After the molecular perception the similarity coefficients are converted into 5 character ASCII codes and combined to a 50 character atom code For all atoms in each molecule being added to the COSMOfrag database CFDB these strings are stored in the CFDB txt file together with a link to the molecule itself Thus the search for most similar atoms can be done by a simple search for most similar atom codes Keeping the CFDB ordered alphanumerically the search for the most similar atom can be done by recursive interval splitting which is very efficient in a large database The time required for finding the most similar atom only increases with the logarithm of the database size In this way for all atoms of a new molecule image atoms of maximum local similarity can be efficiently found in the database Finally those molecules out of the CFDB are chosen for partia
19. ile fdir Sets the directory where to search for the cosmo cos or ccf files of the COSMO calculations In the COSMOtherm compound list the compound indicated by f CFcomp is taken from the structure file list following the COSMOtherm input template An example input file for the calculation of the 1 octanol water partition coefficient is shown in tabular 1 Detailed instruction about COSMOtherm calculations can be found in the COSMOtherm documentation Tabular 1 COSMOfrag input file for the calculation of the 1 octanol water partition coefficient Line Keyword Comment al Action 7 CFDBDIR STRDIR These COSMOfrag Keywords must be in the first three lines STRDIR can be 2 LICENSEDIR omitted if compounds are specified f with their absolute file paths 3 CTCALC CTtotCH NMETA 3 4 start ct input Start of COSMOtherm Inputs This command has to be in line 4 5 cdir ctd fdir This keywords must be in lines 5 and 6 6 7 I logP Comment or empty line 8 f h2o0 cosmo The following lines should be like these Here the partition between water and 9 f l octanol cosmo wet 1 octanol is calculated at a temperature of 25 C 10 f CFcomp B 11 tc 25 logp 1 2 xll 1 0 x12 0 24 0 76 vq 0 11415 wcomp 3 logp amine_corr 0 6 Ved S 12 end ct _ input End of COSMOtherm input template alps smi CC O clecce F ccl Start the compound list OR 13 l
20. iles 8 MOPAC out files also of multiple jobs CFDB_LIST nnn calculation of the first nnn compounds of the CFDB the complete CFDB is calculated by CFDB_LIST Keyword has to be in the structure file list Lists containing files of different formats are permitted Empty rows terminate processing of the list 3 3 COSMOtherm calculations COSMOfrag is a shortcut for high throughput COSMO RS calculations COSMOfrag calculates the o profiles by a composition of partial o profiles taken from locally most similar fragments from the CFDB The time consuming quantum chemical calculation is avoided Generally the whole functionality of COSMOtherm is available and only restricted by the limitations of COSMOfrag e g restricted to neutral molecules However we recommend predicting only those properties which can be well be described on the drug like level 3 3 1 COSMOtherm input section The COSMOtherm calculation is enabled by the keyword CTCALC in the COSMOfrag global command lines The COSMOtherm input template is similar to the COSMOtherm input file Inside of the COSMOfrag input it is enclosed by the two lines Sstart CT_input as the fourth line and Send CT_input as the line closing the COSMOtherm template input section The following parameters have to be specified in the input template 9 ctd named cdt Use the file name ctd as COSMOtherm parameter file cdir Sets the directory where to search for the COSMOtherm parameter f
21. ist list referes to a text file with the name of the compounds The hash character is used to identify comments in the input file If the hash character is the first character of a line the complete line will be ignored In other positions any text after the hash character will be ignored This applies only to the COSMOtherm template 10 3 3 2 Property prediction by COSMOtherm The COSMOfrag program can be used for COSMOtherm batch processing Basically there are two application scenarios e A COSMOfrag run in fragmentation mode ACTION 7 and subsequent COSMOtherm calculation on the produced metafiles A path to the input structures must be given with the keyword STRDIR e COSMOtherm batch processing on existing cosmo ccf mcos files without COSMOfrag database operations ACTION 1 A path to the cosmo ccf mcos files must be given with the keyword STRDIR For the following list of properties a fully automated batch calculation can be performed Here the calculated properties are written in tabulated form to a file with the extension res ready for post processing in a spreadsheet program 1 Water solubilities or other solvents 2 Partition coefficients Octanol Water or any two phases 3 Activity coefficients 4 QSPR properties a Intestinal Absorption Coefficient logKIA b Blood Brain Partitioning Coefficient logBB c Organic Carbon Soil Water Partition Coefficient logkOC d logKHSA
22. l images which have most similar image atoms for a large number of atoms of the molecule to be fragmented Based on our present database of presently more than 100000 molecules and drug like compounds most molecules get fragmented into 2 4 fragments Based on the 7 order similarity coefficients of all atoms a unique coefficient is generated for each molecule This is converted into a 10 letter ASCII code being unique for each molecule apart from cis trans isomerism and stereo chemical differences Cis trans isomerism is translated into a 11 letter and stereochemistry into a 12 letter Thus a 12 letter unique name results for each molecule This name is used for the identification of molecules in the CFDB database and as name for the compressed COSMO files The simple unique molecular names may be useful even in contexts different from COSMOtherm property prediction If favored compounds should not be sufficiently represented in the CFDB you can run DFT COSMO calculations for a small number of representative compounds add them to the database and consecutively improve the fragmentations The COSMOfrag software will assist you in finding appropriate candidates for database enhancement COSMOfrag efficiently performs the fragmentation of a new molecule by molecules of the CFDB The result of the fragmentation is written to a COSMOtherm meta file mcos which then can be used as substitute for the full COSMO file in the COSMOtherm input Thus
23. l molecule and to replace it by a composition of the o profile out of partial o profiles taken from locally most similar fragments of molecules whose DFT COSMO calculations are stored in a database If for all parts of the molecule images with a sufficient degree of local similarity can be found in the database this fragmentation causes only a minimal loss of accuracy Since the fragmentation even for a large database of several ten thousands of molecules requires only parts of a second the total property calculation including the subsequent COSMOtherm calculation takes half a second per compound on average Thus property calculations for about 100000 150000 compounds are feasible within a day on a 3 GHz computer Probably the most essential part of COSMOfrag is a thorough molecular perception of the input molecule which can be given in different molecular formats including SMILES notation In this perception eventually missing hydrogens or bond orders are added Bonds and rings are analyzed with respect to conjugation E Z substitution and aromaticity Finally for each atom hash coefficients are calculated taking into account all the local information about the atom itself and about bonds and neighbour atoms Step by step higher order hash codes are calculated which include the information about an increasing number of neighbouring spheres of the atom Thus two atoms that have an equivalent molecular environment up to the n neighbouring sph
24. nalysis wrong stereo definition or stereo centre with more than 4 neighbors strange hybridization gt Erros in metafile generation too many loops in generation of metafile generation of metafile failed atom not assigned in makemeta gt Errors due to set filters gt 38 gt Invalid structures gt 11 gt 13 gt 23 gt 24 gt 25 gt 26 gt 27 gt 28 gt 29 gt 33 gt 36 gt 14 gt 15 gt 35 gt 34 gt 37 gt Output errors gt 64 number of heavy atoms larger than MAXNUMHEV number of heavy atoms smaller than MINNUMHEV Error in writing ccf file 13 WARNINGS v The structure has no z coordinates Make sure that it is a valid 3D structure gt 2 The structure has unrealistically short bonds This geometry not considered as valid 3D structure v 10 Atrans conformation was found for a carboxylic acid or an ester like fragment This is unlikely 12 Isotope information is ignored 13 Dueto the lack of cis trans information trans was assumed for a double bond 20 Extreme sigma value beyond 3e nm2 found 30 Compound from PRF line processed without structural details 40 Compound most likely will be zwitterionic in water 51 Conflict in SMILES EZ assignment 52 No single bonds in ring Write_ SMILES failed 53 No valid ring assignment found within 20 attempts 54 More than 99 rings Write_SMILES failed 60 Erroneously identical UNIQUE NAME found 87 93 Molecules has cha
25. on the list generate a unique SMILES string with chirality information for each file on the list generate a unique SMILES string with EZ information for each file on the list trigges writing of SMILES for all possible EZ isomers to the out file write descriptors to the tab file no of atoms hydrogens bonds ringbonds conjugated atoms internal hydrogen bonds rotatable bonds no of terminal alkyl carbons no of terminal alkyl groups molecular weight no of atoms in of the largest ring LargeRing generate car for each 3D file on the list generate xyz for each 3D file on the list generate sdf for each 3D file on the list generate ccf for each cosmo file on the list generate a _mcos car file for each successfully fragmented file on the list that visualizes the fragmentation 3D input structures will be displayed central ACTION and CFDBDIR are indispensable for COSMOfrag runs If STRDIR is omitted the absolute path for the compounds can should be specified 3 2 2 Structure file list COSMOfrag supports the following structure file formats 1 SMILES a Direct input via list Notation smi CCC OCC b SMILES in smi file s Multiple smiles strings in one file supported Names after Smiles supported 2 SDF MDL Isis a Single or multiple sdf files 3 CAR MSI Biosym 4 XYZ 5 COSMO Turbomole CCF compressed COSMO COS Mopac COSMO 6 ML2 Sybyl 7 MOPAC dat mop arc or arcxx f
26. rge iwrn 90 iwrn 90 means zwitterion VV VV VV VV VV WV 4 The COSMOfrag database The COSMOfrag database CFDB of the current release consists of 112371 common compounds and drug like structures collected from databases like Physprop NCI NCBI PUBCHEM and various others The quality of the COSMO RS property predictions resulting from COSMOfrag metafiles depends on the quality and size of the CFDB database The database molecules have been calculated with TURBOMOLE on the BP SVP AM1 level i e by a single point DFT calculation with SVP basis set on a molecular geometry optimized with AM1 COSMO with a modified MOPAC 7 program In the majority of the cases the initial 3D structures are generated by the CORINA program by Molecular Networks It is recommended to use compounds calculated by a comparable procedure for addition to the CFDB Physical Properties Database by Syracuse Research Corporation National Cancer Institute Database National Center for Biotechnology Information 14 5 Applications 5 1 Visualization of metafiles To visualize a fragmentation corresponding to a certain metafile setting the keyword WMCOSCAR will generate a structure file in CAR format which contains each CFDB molecule occurring in the metafile For a 3D input structure the input molecule is placed in the centre of the visualization Opening this xx_ mcos car file in a molecular viewer the fragmentation is comfortably visualized and b
27. sing of the input 3 5 COSMOfrag output The output of each COSMOfrag run is stored in filename out and filename tab where filename inp is the name of the input file For named metafiles for the structure file list the named option can be added to any keyword ACTION 2 7 Results of COSMOtherm calcultations for all compounds in the structure file list are written to a tabulated output file filename res 3 5 1 Metafile quality Even though in general an evaluation of the quality of a specific fragmentation is not possible the total COSMO charge can provide an indication of low quality metafiles By setting the CTtotCH keyword the total COSMO charge will be tabulated in the output file filename res file Values gt 0 4 or lt 0 4 indicate possibly bad fragmentations and should be separated out although in some cases even those metafiles can produce good property predictions due to error compensation 3 5 2 Errors and warnings Due to its functional range COSMOfrag offers a rather long and detailed list of error codes These errors predominantly occur within the structure parsing part of the program and are listed in filename tab separately for each molecule on the list Some of the error codes are of interest only for the development of the program In the following list we have grouped the error codes corresponding to their causes They should help the user to filter wrong structures or structures not supported by COSMOfrag ERRORS
28. the path is set correctly you should be able to call COSMOfrag on the dos shell from any directory by typing cf exe name inp where name inp is the name of the input file which should be located in the working directory If the path for the COSMOfrag executable is not set you can still run a COSMOfrag calculation by typing in the absolute path for the executable e g C Programme COSMOfrag cf exe name inp 3 Running COSMOfrag 3 1 Program execution COSMOfrag is started from a Linux shell with the following command cf lt input file name inp gt And from a Windows shell with cf exe lt input file name inp gt COSMOfrag requires an input file with the extension inp e g filename inp The output of each COSMOfrag run is stored in filename out and in a more comprehensive way in fi lename tab Results of COSMOtherm calculations are written to a tabulated output file filename res 3 2 Input File The COSMOfrag program requires a formatted input file consisting of the following parts e Three global command lines e Optional COSMOtherm template input e List of structure files or smiles strings 3 2 1 Global command lines All commands are given in the form command or command argument or command argl arg2 i e if several arguments are given for a command the arguments have to be included into curved brackets and separated by blank spaces Commands are not case sensitive Of course paths and filenames on Unix or
29. y displaying the name field first column the weights of the metafile can be displayed 5 2 Expansion of the COSMOfrag Database CFDB Although the CFDB now consists of more than 100 000 drug like compounds it is possible that a few molecules out of a larger set or groups of certain functionality are not represented reasonably To identify those compounds the MAXSTRING keyword can be used A string in which the maximum similarity of the atoms is given in input order is then written to the output and table files From our point of view a similarity value 1 can always be regarded as adequate O similarities on the other hand should be replaced in either case COSMOfrag therefore denotes these molecules with error code 38 Concerning 1 similarities no definite advice can be given For fast screenings predictions with a moderate loss of accuracy may be acceptable and the number of compounds with 1 similarity can be quite high For high quality predictions with a small number of T similarities one could think of replacing them by molecules covering the missing fragments A pre screening for compounds with low similarity can be done by the following procedure e Run COSMOfrag with ACTION 8 on your set of molecules In this run molecules which cannot be fragmented adequately with molecules from the CFDB are added virtually That means the hash coefficients for the molecule are generated and added to the database in th
Download Pdf Manuals
Related Search