CzeekD Manual
Contents
1. Figure 10 b The number of bonding points as well as the number of R groups must also be defined Setting the R groups can be done using the R1 to R4 buttons in Marvin Sketch Structures that are created using other tools can also imported from files and by copying and pasting from other applications Please refer to the following link for details regarding the use of Marvin Sketch http www chemaxon com marvin help sketch sketch index html In our example the unit at the middle will be set to the structure of Quinazoline as shown in the Figure below File Edit View Insert Atom Bond Structuri Calculation Tools Help 0 o 8 b Figure 10 Drawing the core structure 3 3 2 Selection of fragment library There are two ways to select the desired fragment library 1 by right clicking on a unit then clicking on Select library item on the context menu and 2 by selecting a unit by clicking on it once then clicking on the Select library button This will open a window that will allow the selection of one or more desired fragment library from a list Several fragment libraries are available depending on the structure and intended usage Please refer to Table 2 for details The number of fragments contained within the selected fragment library can still be narrowed down by using one or more methods shown below e According to suppliers this is possible only with those libraries containing supplier information e According to qu2. the distribution map is presented as the plot of molecular weight at the X axis Weight against the score at the Y axis Changing the parameters plotted in any of the axes can be done by clicking on any of the axis labels then selecting the desired parameters in a popup window Figure 19 a In this example CGBVS scores of GAK X axis and EGFR Y axis are plotted CzeekD GEG cos ee I I e n I DO gt eee lt HH Figure 18 Results analysis window Figure 19 a Axis display parameters and b Distribution map Kyoto Constella Technologies 13 CzeekD Manual Clicking and dragging on an area within the distribution map will select the compound present in that area Clicking the List view button will show the compound structures with its corresponding scores as well as other parameters Figure 20 Selecting a range by specifying values or selecting all the plotted compounds can be performed by clicking on the Range or Select all buttons respectively Canceling the selected range can be done by double clicking within the distribution map The maximum number of compounds that can be selected is 50 000 Czeek N a cartins are E ian te nce lane crm nmin dv tener o A dwn 1 Click and drag to select an area t_ Double clicking within the map will reset selection Nn Click List view button Click on the Numeric button to do numeric search Hebe Figure 20 Selecting a
3. of fragment display in the distribution map the X axis is initially set to freq that indicates the frequency in which the fragment is utilized in the generated compounds The Y axis is set to mean that indicates the average score of compound groups possessing that fragment In addition fragments that have available supplier data also show the name and ID of the supplier That is true in the case of building block library being offered by Namiki Shoji Namiki BB Library Kyoto Constella Technologies 17 CzeekD Manual Czeek D Ce oo rece coi rm try opto du r tved cea tee 7 eat os D n Eu 4 04 m 03 a i t TE FF E Ti E ARTIF TEEST EEr M X l hei f gt Wium Le rt o a Saen anni e sz mR ti i4 na Cee soot ness ua QUEST I WA ete RI LI owe le Co Pena a L 1 PEM sana di 21212424 lt lt lt T lt P E i i it Figure 26 Displaying structures of selected fragments 4 5 Refining compound search Normally CzeekD generates several hundred thousands of compounds during calculation While it depends on the selected target and synthesis frame settings the number of high scoring compounds can reach up to tens of thousands In our example calculation using EGFR and GAK alone there are 21256 compounds scoring 0 8 or greater and checking that list for compounds worthy of further examination would be very difficult if not im
4. setting window In this example the structure of Gefitinib is drawn while EGFR and GAK were set as the target and off target proteins respectively Figure 34 Kyoto Constella Technologies ae CzeekD Manual Czeek D SS jetta o Peoficdm Lint Search 2 Ure Camo Catan Carp Ara range Tama striitare Dese ve PIT e Gt bore ne 3 Draw structure ti sam asce 4 erat nre lt IM Inuteihe g taf A H tnt dii n y srein Me Figure 34 Calculation based on structure entered using the Marvin Sketch plugin The result of the above calculation is shown in Figure 35 una 44624404 COP COORD ddt 447 2 dit OEP det ont aay l Puen lone gee SS ieee Figure 35 Scoring result for Gefitinib 5 2 2 Uploading structure files If calculation of a set of compounds is desired the compound structures must be put in a file SD SMILES or mol format and then uploaded into the system Up to 100 compound structures can be uploaded The first line of the descriptors for all compounds must start with the compound name and unexpected results will be obtained if that information is missing A unit without any bonding point must be placed first in the synthesis frame area The unit is then selected before clicking on the Select library button A window similar to that in Figure 36 will open Click on the Browse button to select the file for uploading The number of compound structures
5. showing the information for the current user Figure 32 b Information can be modified by clicking the Edit account details button This will open a dialog box that allows the user to change the user name e mail address and password Figure 33 Click on OK to close the window and effect the changes or Cancel to cancel current operation tam informa bor it Welcome to CzeekD di User informatica Gau Click to check user information a Click to edit user information b Figure 32 Displaying user information Edit user information x Usemame E mail taylo k ct jp Password Re enter password Cancel Figure 33 Editing user information 5 2 Compound scoring by CGBVS With CzeekD scoring of a predefined structure against one or more protein targets can be performed and the structure can be specified in 2 ways 1 drawing or uploading the desired structure using the Marvin Sketch plugin or 2 by uploading SD or SMILES file within the fragment library selection window 5 2 1 Using the Marvin Sketch plugin First remove the default frame by clicking on the Clear frame button From the selection of units select the one without any bonding points then click on the frame area to place the unit Select the unit by clicking on it then click on the Draw structure button to activate the Marvin Sketch plugin Draw the desired structure then click OK to return to frame
6. Kyoto Constella Technologies Co Ltd CzeekD Manual December 2014 CzeekD Manual TABLE OF CONTENTS L PAREN ONC NO ON PEPPE CREO RR RO RR ORE 1 Lil IR it MM SC a E EE EE TE E suacocsusancdmacseass l L2 Logn andoperaional Pollyanna rain l 2 Creation of Compounds in CzeekD iscsiccsiticioiaricinicnaneaniana ia lcisivadinonreiniadbisasibciilanibviatiaialasegibciniazzbinri ani 2 21 COMPOUNAEFANOMSCHEIO iridata arreca ode 2 2 2 Retrosynthetic Combinatorial Analysis Procedure RECAP 2 2 3 Particle Swarm Optimization PSO Algorithm ceccsssceccccceeceseseeeeeccceeeecaaeseecesseeeeceesessaaeeeeees 3 3 Operation Specifying Calculation Settings and Actual Calculation vcrescrrrrrecereeeeneeecececeseeeeeeeere 4 dl BOI 5 gt IIS dae 7 AI RR A 7 E Bi nia TN 9 3 5 Regarding Combinations Chemical Space and Estimated Time i 11 3 6 Checking the status of a calculation JOD 12 A Calculation Results Analysis ccccscssscesteessssvscessavetassacseveseisenusseasecensnesvenvevesstaseantesseuensessenveseaveasioanesessassonens 13 Al Digione 13 42 ompound HL ope 110 9 innata TES 14 43 Nu mericandpariialstr tur SEARCH E a N 15 4 4 Compound display and fragment display I7 460 Liliana 21 S I I RCA RR RE RR RESA 22 5 1 Modifying user information changing passWord 22 52 COMPOLAd scornet by COBY S serseri
7. antitative physical properties number of heavy atoms molecular weight and ClogP e According to partial structure can be done by fragment structure and structure similarity search Kyoto Constella Technologies 8 CzeekD Manual After the conditions above have been set clicking on the Search button will start the search for fragments fitting the set conditions Clicking on the Show Fragment List button will reveal the selected fragments on the right side of the window Figure 11 a Figure 11 Fragment library selection window The example in this section sets the left unit to RECAP non aro ring 1 and the right unit to RECAP aro ring 1 fragment libraries RECAP non aro ring ge3 absent Table 2 Fragment library list RECAP chain ge6 Has 6 or more rotatable bonds but no ring structures 3 4 Bond settings In CzeekD the RECAP rule is applied to bonding between fragment units and the fragment that conform to these rule are automatically selected However if a particular structure is set for a unit the RECAP rule that must govern the binding must be explicitly defined The procedure is explained below As shown in Figure 12 a the bond to be set is selected by clicking on it once after which the Set bond rules button 1s clicked This will reveal a pull down menu from which the appropriate R group is selected A window similar to that in Figure 12 b and 13 will then be shown At the upper pa
8. atus is usually shown in the following order Queued Pre Processing Running shown as percentage Post Processing Normal end Results of complete jobs can be displayed and further analyzed A currently running job can be aborted by clicking on the STOP button which is displayed in place of the trash bin icon in case of a currently running job Figure 17 Job list window The trash bin icon appears in completed jobs as well as in aborted jobs Clicking on the trash bin icon will delete the job Deleting a job cannot be undone so it is advisable to download your results before doing anything else so as not to accidentally lose important data Kyoto Constella Technologies 12 CzeekD Manual 4 Calculation Results Analysis In this section display of results and analysis feature will be described We have used EGFR and GAK as example targets since section and we will continue to use the same in this section Double clicking on a job or clicking on the red icon at the rightmost column of the job list window will display the results of a calculation job 4 1 Distribution map A sample of the results analysis window is shown in Figure 18 below The left hand side shows the data manipulation area while the right hand side displays the selected compound structures with corresponding scores and parameters Compounds shown in the distribution map range in colors from blue to nearly read indicating the density number of compounds Initially
9. based fragment search refinement In compound display mode select an area by clicking and dragging within the distribution map Click on the List view button to show the compound structures then select one or more compounds by clicking on the checkboxes Figure 29 On the Frame panel click on the left or right fragment units The distribution map will change to show only the fragments present in the selected compounds Figure 30 Kyoto Constella Technologies 19 CzeekD Manual CzeekD GER cs Wweicome to czasy be etnea Peek ieee Oe Soe ipo Seed eworciws Bhs wget ee Dapy ptor Hee emrt beas Magnaat d rect were Camera COR PAPMAN GAK IRAAN Ac regno Wage 9 a 44 d Osy Ubi 1 4 TPA 377 2 440 374 000 FIT 73328294 Wo 0 Oe 01039 Magna te Wm Sewa Oa a Linee teint at Setociot 15467 Figure 29 Selecting compounds Before search refinement After search refinement 104 91370 9 4927 80836 04870 8964 77 48 1312 1147696406 154 919320 9 492 80030 84870 8984 71148 1352 14 761640 8 fr g treg J seoce TP numere Jo ne teamed Ri aR St TI Cl Aste be ua n di f 0 Figure 30 Change in distribution map after compound based search refinement For both cases above the results of the refined search can still be narrowed down further by setting different conditions and by clicking the checkboxes to select fragments or compounds Clicking the Save button located below the distributio
10. contained in the file will be shown at the lower left hand area of the window Clicking the Show Fragment List button to show the structures at the right Kyoto Constella Technologies 23 CzeekD Manual side of the window Click the OK button to set the compounds for calculation is Uptoed Me moU SOSMES r t List Select file EGFRI smy write ID HA Ma oy sopies Sabi homn Eaa y 377063 ji da i Library List z ded a von 1 Click to select file to upload N 2377104 HO nr 417387 O OS Supplier List ven Lul Soori Fragmants 3122 ALA g Ci Pa 412200 i ri UL r Refine Search E w 2 Click to check structures Fy uw lt eavy Morn lt hi F p x Ostructuce timer Y lag O 1 Ox Cancel Sa pou as x ami fy Figure 36 Uploading structure via the fragment library selection window Results for the above calculation is shown in Figure 37 below CzeekD ZEMmDE ca racerne te Conway dine eteri GEER H tz 423 U tr deed vespri eee es Oo Suor avanzo ae eee x Py mora gt n v artn er n O del Rod DOC 4 Bag 3 ETTI AIR dll ARA TARSIA dr al r cay nr me ae 2 Lt e na o Titas Seat o ee TI Te bc 5 3 Assigning s ry a 4 s i i us fa Figure 37 Calculation results for compound structures uploaded from file everal core structures Several core structures can be assigned using the Marvin Sketch drawing tool As shown in Figure 38 it can be done by
11. culation is currently being shared with everybody within the same group Clicking on the same icon will un share the data Only the owner of the calculation results can share and un share the data Unchecking the Show shared data checkbox will reset the job list to display only the currently logged in user s data Checking the box again will show currently shared data by other group members Kyoto Constella Technologies ZO CzeekD Manual Aga mew jod update Search Share Lise lob name Gian Chart Display options Vi Show shared data icon desonpbon ne Eno bme Running time Expected compton bme One ar on p test kinase egfr TETY 2014 8 21 11 30 2014 8 21 19 15 07 44 57 2014 5 21 18 53 ww ti teat kinase egfr i mary 407 4 6 22 10 42 eVl4i igs 14 54 04 12 32 2014 G 22 14 27 E G Taviofdmund EGFR GAK Testi Normas end 2034 32 16 10 25 2014 12 16 18 13 07 38 13 2014 12 16 18 09 G amp G g Go T Todo Edmund Geftinb EGFR GAK Test Norma end 2014 12 18 4 12 15 1 VOl o G T Normaond 2014 12 22 16 58 2014 12 22 10 39 00 91 07 G O Normans 2034 32 22 17134 2014 12 22 17 35 00 00 57 O gG O 2 Changes to red color indicating shared status 1 Click here to share data for this job Figure 39 Sharing calculation data 5 5 Expanding the range of compounds that can be generated The synthesis frame can be designed in such way that the number of virtual compounds the chemical space that can be generated will increase As shown in Figu
12. d by the following equation X Gia Fy In the above equation x is a vector that represents the properties of the fragment while u represents the unit Since a virtual compound is created by combining all the fragment units in the synthesis frame the position vector X can be represented by all the fragment vectors x connected in series Although the basic equation for PSO indicates that the position vectors are continuously changing with regards to de novo design the structure of the compound must be obtained from the value of the position vector X The approach that is used to realize that is to select all the fragments with vectors whose properties are very much alike with those of x that constitutes X and subsequently bind them to create the virtual compounds If several proteins are selected an overall score will be calculated Score calculation is represented by the following equation s f P positive i 1 P negative In the equation above n represents the number of proteins S represents the score of the protein and if negative that is off target protein is selected S is calculated using 1 P 3 Operation Specifying Calculation Settings and Actual Calculation In this section de novo design settings will be explained using Kinase as an example target The target proteins are EGFR and GAK Compounds that are active against EGFR while at the same time inactive against GAK will be Kyoto Const
13. drawin g several structures within Marvin Sketch Figure 38 a Structures not connected together by Kyoto Constella Technologies 24 CzeekD Manual bonds are recognized as independent structures Figure 38 b There are two important points to remember when using this method number of R groups should be the same for all the structures 2 R group symbols should be the same for all the structures For the second condition above if 2 R groups are present the R1 and R2 symbols must be used for all the structures and not R1 R2 for one then R3 R4 or R5 R6 and so on for the succeeding structures Failure to follow the conditions above will lead to unexpected results Please note the system does not give a warning if you commit such errors and will accept the invalid settings and continue on to calculation aon Marvin Sketch G hites denovo czeek jp creekd Ma File Edit View Insert Atom Bond Structure Calculation Tools Help vr 2 a cy t A b Figure 38 Drawing more than one structure within the Marvin Sketch plugin 5 4 Sharing calculation results Users belonging to the same group defined during creation of accounts are allowed to share their calculation results with each other This can be easily done by clicking on the purple icon containing white callout symbols Figure 39 This will make the gray button located at the leftmost column to turn to red color indicating that the results for that particularly cal
14. ed not to drastically increase the number of iterations If a much broader search range is required performing the same calculation several times is a better option You can however make changes to the settings when the need arises 2point Figure 42 Optimization algorithm settings Kyoto Constella Technologies 28
15. elected and these proteins will be shown on the lower right hand panel Figure 7 In the case of Figure 7 EGFR will be the target and GAK will be the off target Kyoto Constella Technologies 6 CzeekD Manual 3 2 Synthesis frame settings Setting the required conditions for the creation of virtual compounds is performed in the synthesis frame panel Operation for the required settings shown below introduced in the section 2 will be explained in this section set number and configuration of units topology 2 specify structures and or fragment libraries to be used 3 select rules that would govern binding between units Immediately after logging in to CzeekD the settings window will be shown Figure 6 wherein a synthesis frame with default topology 3 units arranged horizontally is initially set In our example this synthesis frame can be used without changing the topology But if the topology has to be modified the default synthesis frame must be deleted first by clicking on the Clear frame button Creating a new synthesis frame can be performed by first clicking on any of the desired unit located on the right side area of the Add units label upper left hand area of the panel Figure 8 Clicking on the frame area will create the selected unit After placing the desired units connect the units by clicking and dragging between the connection points of each unit There are 5 types of units that can be used in CzeekD each
16. ella Technologies 4 CzeekD Manual calculated searched for this particular example As shown in Figure 5 compounds will be designed having the Quinazoline part of Gefitinib as the core structure with both ends serving as fragment attachment sites Attachment site 2 Attachment site 1 Gefitinib core structure Figure 5 Compound design scheme 3 1 Target selection There are 3 general settings that must be performed by the user a select target protein calculation model b specify synthesis frame settings and c set calculation parameters The last is optional and default settings will be assumed if not set by the user The window below will be shown immediately right after logging in to CzeekD On the left side of the window is a panel that allows the user to select the target protein family and the actual target proteins itself The biggest panel located at the right side of the window allows the setting of the synthesis frame Below is where calculation conditions can be set and is also where the Start job button is located CzeekD GE mie o caan rr ES Men TE we U O w wa ape of ime toe rar Ln Seta fe Lia all t det te Figure 6 Settings window In selecting the target protein the CGBVS calculation model should be selected first Figure 7 Models can be selected after clicking on the dropdown list There are 6 standard models to choose from Table 1 Once a model is selected by clicking
17. est solution is the unique feature of PSO The general equation for the PSO algorithm is shown below Each particle has its own position X and speed V vectors and based on the equation searches the chemical space for the best solution Xa X V V1 WV tne Care rcz Came X In the equation above Monee refers to the best solution found by each particle personal best X gbest refers to the best solution found by the entire swarm global best w refers to the inertia constant r r indicates random digits from 0 to 1 while c4 c gt refers to the size of particles attracted to the best solution The position and speed vectors are constantly recalculated and changes in their values lead the swarm nearer to the best solution There have been other equations that are proposed to represent PSO but in our case we are adopting the equation above If you would apply PSO to de novo design a particle represents a compound On the other hand a particle s position vector X is a value that represents the properties of the compound represented by the particle The fitness function score that should be optimized is one that is independent of X and in the case of CzeekD the CGBVS Chemical Genomics Based Virtual Screening score is adopted This score is also the same score calculated when using our other product CzeekS when several proteins are specified the average score is presented The position vector X is represente
18. having from 0 to 4 bonding points Figure 9 If a unit having no bonding point is selected a simple scoring screening calculation will be performed refer to p 22 for details There is a limit of 9 units that can be used to compose a synthesis frame Furthermore one or more units that have one or more unattached binding points are not allowed CzeekD GG eee uz Select unit M tom dere Place unit ona Connect units Figure 8 Selecting arranging and connecting units SE Figure 9 5 types of units according to the number of bonding points 3 3 Unit settings Unit properties can be set in two ways 1 drawing a structure through the use of a plugin integrated within CzeekD and 2 selecting a fragment library 3 3 1 Drawing a structure With this method a structure fragment is set by using a drawing tool integrated within CzeekD As shown in Kyoto Constella Technologies CzeekD Manual Figure 10 a it can be done by right clicking on one of the units then clicking on the Draw structure item on the context menu The Marvin Sketch Figure 10 c plugin will appear in a popup window that would allow the user to draw the desired structure Another way to activate the drawing tool is to select a unit by clicking on it once then clicking on the Draw structure button Clicking on the OK button after drawing the structure will close the plugin and return to the CzeekD window already showing the structure drawn
19. ments referred to in CzeekD are basically structures created after cutting compounds using the RECAP method There are actually 11 basic rules that are followed in the RECAP method and these are indicated in Figure Kyoto Constella Technologies 2 CzeekD Manual 3 below 6 Ce NNI S O Za ii A Nd pm h Pm O Di Pr Am we 7 Th e i opt we omy k a a Al N 8 Ul 1 amide 2 ester 3 amine 4 urea Y ar b Ly 4 eu Eu a illa Np N Np HEX buy n 4 oe Gd H Mal Bk pP I Sig af N Sa a i lea Dal 3 La Pu Pu ty VA N 5 ether 6 olefin 7 quaternary nirogen 8 aromatic N carbon K K Vi O l e ai I P U i i Il H s i n Fy Yb N x Lu ur a Fd gt l r I 0 O S 0 5 9 lactam N carbon 10 aromatic carbon aromatic carbon 11 sulphonamide Figure 3 RECAP rules In CzeekD we employ heavy metal symbols such as rare earth elements as markers for cutting and binding sites These metals are not normally used in the creation of drugs therefore are used only as markers Markers differ based on the cutting rule and the type of bond created after cutting We also refer to these heavy metal symbols as attachments For example if an amide bond is cut La is attached to the acyl side and Ce is attached to the amine side please refer to Figure 3 for details In CzeekD fragment libraries refer to the collection of structure
20. n area within the distribution map and displaying the structures 4 2 Compound list operations Putting the mouse over a structure shown on the compound list will magnify the structure and clicking on the structure itself will open a structure viewer displaying the clicked structure Expanding the width of the structure column will also result to the magnification of the displayed structures Unchecking the Show structures checkbox will hide the compound structures while clicking the Show fragments checkbox will show the fragment structures Clicking on the column name of any parameter will reveal a dialog box displaying sorting and filtering options Clicking and dragging on the column name of any parameter will change the order of the columns Figure 21 Compounds can be selected by clicking on the checkbox to the left of the compound structures Selected compounds can then be exported in SDF or CSV format by clicking on the File export button Kyoto Constella Technologies 14 CzeekD Manual a b Figure 21 Compound list panel 4 3 Numeric and partial structure search Display filtering has been described in the previous section but actually in the first place only 50 000 compounds can be displayed However compounds to be displayed can be narrowed down by filling in search conditions for the distribution map There are two ways that this can be done first by using numerical parameters and second by partial st
21. n map will save the conditions and results of the current search Up to 3 searches can be saved and you can switch between them by clicking on the buttons x1 x2 or x3 buttons located above the upper right hand corner of the distribution map Saving more than 3 searches will Kyoto Constella Technologies 20 CzeekD Manual erase the least recent search Clicking on the Reset map button will reset the distribution map to display the default and will erase all the saved searches 4 6 File export Data of the selected compounds can be downloaded by clicking on the File export button and selecting the appropriate file format within the dialog box that appears Figure 31 Data can be saved in either SD or CSV file format Structures saved in CSV files are in SMILES format Clicking on the OK button allow the user to specify a filename before saving the file File Format Please select export file format Export SD file Export CSV file OK Cancel Figure 31 Selecting file format before saving data file Kyoto Constella Technologies 21 CzeekD Manual 5 Other Features This section describes the other functions of the CzeekD system not described in the previous sections 5 1 Modifying user information changing password Registered name e mail address and passwords can be changed by first clicking on the User information button Figure 32 a The display will switch to a new window
22. n scheme Virtual compounds are created according to the conditions that are defined in a synthesis frame Figure 1 The synthesis frame is generally composed of units and bonds Units represent areas where fragments are placed whereas bonds are parts where binding rules between fragments can be defined Appropriate fragments are placed into all the units and results to creation of a virtual compound after binding of the fragments involved Structures that do not conform to the defined conditions of the synthesis are not created The following should be defined in the synthesis frame number and configuration of units topology 2 structures and or fragment libraries to be used rules that would govern binding between units Lb QAL se O Figure 1 Example of a synthesis frame There are 5 types of units Figure 2 that are available in CzeekD each differing only in the number of bonds possible In creating the synthesis frame units are selected and arranged then are bound according to the number of bonds available for each unit Units with dangling bonds are not allowed Up to 9 units can be set in a synthesis frame Figure 2 5 types of units 2 2 Retrosynthetic Combinatorial Analysis Procedure RECAP Fragments libraries provided in CzeekD are created based on the RECAP rules Using the RECAP method a particular compound is fragmented cut based on the rules by which it was originally chemically synthesized Frag
23. on it available target proteins will then be shown in a tabulated list Figure 7 The Ligands Kyoto Constella Technologies D CzeekD Manual column shows the number of compounds used in machine learning for each target protein and the Compound Clusters column indicates the number of structural variations of those compounds Please keep in mind that in CGBVS higher number of structural variations between compound used in machine learning equates to higher prediction accuracy Standard model for Nuclear receptors Standard model for Proteases Standard model for Transporters Table 1 List of standard models Optimizations PoC Ligands 255158 Coren eens Figure 7 Protein selection panel Figure 7 shows an example in which the Kinase standard was selected A protein can be selected by clicking on the checkbox located on the left side of the protein ID Clicking on the box once will show A and a second click will show W A third click will return it to an empty checkbox A indicates that active compounds against the selected protein will be designed while V indicates that the compounds to be designed should not be active against the selected protein In short A and V indicate the target and off target respectively The protein IDs are hyperlinks to the UniProt database and clicking on them will open the web browser and show the page containing information about the clicked protein Up to 10 proteins can be s
24. orm de novo design calculation mati TE Optimization PSO Advanced Car oll Al EGFR E Estimated time 43 00 Ligands 755128 srt Combinations L S410 Figure 15 Summary of calculation settings Combinations refers to the total number of compounds that will be created based on the synthesis frame settings It is basically shown as a high value exponential notation Ligands refer to the default number of compounds actually included in the calculation This value depends on the number of particles and number of iterations set within the PSO Advanced settings Estimated time indicates the approximate time it will take for the calculation to finish and is meant only as a guide l1 Kyoto Constella Technologies CzeekD Manual After carefully checking the settings the calculation can be commenced by clicking on the Start job button A dialog box just like in Figure 16 will be shown and a job name for the calculation can be entered As much as possible use an easy to comprehend job name Job Name Input Job name Figure 16 Job name input dialog box 3 6 Checking the calculation job Status of calculation jobs can be checked by clicking on the Job List Window button Figure 17 Check the Status column to see whether your job is currently running or have already completed The status of a currently running job is usually shown as the percentage of the completed calculation Job st
25. possible In cases like these refining compound search is advisable and is a feature included in CzeekD 4 5 1 Refining fragment based compound search This is done by first switching to fragment display mode As shown in Figure 27 select an area in the distribution map and display the list at the right panel Select one or more fragments from the list by clicking on the checkbox to the left of each structure All the fragments in the list can be selected by clicking on the checkbox located at the left of structure column heading Next go to the Frame panel at the lower left hand side of the window then click on the Select entire frame button The distribution will then change into something similar to that shown in Figure 28 In this example the fragments having the highest average score mean gt 0 7 are selected so the center of the distribution map would be composed of those fragments Kyoto Constella Technologies 18 CzeekD Manual CzeekD Gaga Co ca Vulcan tp Cot ioe vee GER Figure 27 Selecting fragments Before search refinement After search refinement PIG 377 P4484 624 SOOCR ETT 763 2 820 4006 8O0 B O88 TRO 372244064524 66006 GT 7532 629 4 965 6 93181338 Meg Wepre 1E 1 beso i t mos Oe ty Nurmenc Sruct Cen be savod 2 Namor sn gt can te save Manat map DI Range See si CI Range Selected Figure 28 Change in distribution map after fragment based search refinement 4 5 2 Compound
26. re 40 adding a unit to the synthesis frame will expand the calculation at the R1 side In this case the chemical space increased from 1 83 x 10 to 1 96 x 10 and so is the variation of the chemical structures However increasing the number of units will also increase the time required to perform the calculation LD RECAP_ROM aro_nN Vi i ie RECAP_aro_nng_1 Total 53058 ko PP Tota 194590 Add a unit to the left side of the core structure I RECAP chain les ie tlie Tie RECAP aro nng 1 Fotal 15985 i f a Tota 194580 RECAP nori aro_mn Total 53088 Figure 40 Extending the synthesis frame 5 6 Using custom made fragments As described in section 5 2 structures can be uploaded by using the SD MILES mol file upload feature The same method can be used to upload user designed custom fragments However this type of usage requires understanding of the RECAP rules Kyoto Constella Technologies 26 CzeekD Manual As discussed in section 2 fragments provided in CzeekD are labeled with heavy metal element symbols that serve as markers for proper binding based on RECAP rules These markers are generally removed during the binding process The custom fragments must therefore contain these same markers before being uploaded into the CzeekD system Additionally the first line of information for every compound present in the SD file should contain the fragment name written with a maximum of 10 letters This i
27. re the search PIG 377 7 Atha 624 SHOR G77 7683 8220 4 OF ba rie 1 humen gt Struct Sane ani P ai N e VEO TI kange Select ai Selected 0 ate j o ty 72 2 446 4624 60006 877 7532 8254 06 6 981 B10S8 Wi ceghe P one ce Snxt di Sove E cas be sane Roat mag Lance Sasa 3 Selected Search After the search Figure 23 Transformation of the distribution map after partial structure search Kyoto Constella Technologies 16 CzeekD Manual Figure 24 Resulting compound list after partial structure search 4 4 Compound display and fragment display The Frame panel located at the lower left hand area of the results analysis window displays the synthesis frame Figure 25 At this area clicking on any of the fragments then on the Select entire frame button toggles between compound list display and fragment list display in the Compound List panel Click on any unit to display fragments Click to display compounds Figure 25 The Frame panel Figure 26 displays the fragment structures with the corresponding parameters after selecting the rightmost fragment unit within the synthesis fame then selecting an area within the distribution map In CzeekD the binding point in a fragment structure is designated by an atom of a heavy element not generally used in the creation of drugs Parameters displayed are values from which the effects of the heavy atoms were removed In the case
28. rr E S 22 5 3 Assigning several core Structures tira 24 54A Spanair 25 5 5 Expanding the range of compounds that can be generated 26 5 0 Using c stom made TAC rali 26 5 7 AdvanCed CaleulaviOn Parameters srana a a a a k a aa 2 Kyoto Constella Technologies 1 CzeekD Manual Trademarks All the company and product names appearing in this manual are trademarks or registered trademarks of the respective companies Furthermore trademarks are not appended to all the software and product names described in this manual 2012 Kyoto Constella Technologies Co Ltd Reproduction in whole or in part of this document is strictly prohibited Kyoto Constella Technologies 11 CzeekD Manual 1 Introduction 1 1 Regarding use of CzeekD This manual explains how a computational chemistry personnel can utilize CzeekD with particular emphasis to medicinal chemists Starting from the introduction of the calculation method involved in CzeekD description of operations from calculation up to analysis of results will be described while at the same time showing relevant CzeekD windows It is assumed that CzeekD will be used in a Windows environment and explanations in this manual are based on that assumption CzeekD is a Client server type of web application offered as a service that designs and analyzes new compound structures based on CGBVS Chemical Genomics Based Virtual Screening prediction score Computational servers are main
29. rt of the window the previously drawn structure will be shown with the selected R group as shaded 11 types of bonding rules are shown in Figure 14 and Kyoto Constella Technologies 9 CzeekD Manual with the exception of olefins and aromatic carbons variations are shown in pairs One or more bond types can be selected asa unite O 2 Click on Set bond rules button ton Setect Sbrary Draw stricture Delete unt Edit bomds Set bond rules St bond for IRL Set bond for R2 ar mm RECAP _non aro_rin Li a n RECAP_aro_nng_1 Total 53058 k Pi Total 194560 r n T Select bond a Figure 12 Bond rule settings window Select bonding rule D Gord nse description Nond types Bora cer ka AN e Aio amne z SN A i gin dita cune e ANY de OK Gancqui Figure 13 Setting bond rules for R2 Kyoto Constella Technologies 10 CzeekD Manual m sa amico i ii al Figure 14 Possible settings based on RECAP rules gn 2 lactam N carbon suiphonamide 3 5 Regarding Combinations Chemical Space and Estimated Time The settings required to start a de novo design Job is almost complete at this point Clicking on the accept changes button will make the settings take effect and if there are no errors encountered will give values for the Combinations equivalent to size of chemical space and Estimated Time approximate time required to perf
30. ructure search Clicking on the Numeric button will open a dialog box that allows entry of numerical values that be will be used for the numeric search Figure 22 In this example scores of 0 5 and above and molecular weights of 500 or less were used as numeric search parameters Executing the search changes the display of the distribution map depending on the conditions set in the numeric search parameters Compound search based on partial structures can be performed by clicking on the Struct button then clicking within a dialog box that appears afterwards This opens the Marvin Sketch plugin that allows entry of the desired structure Clicking OK after drawing the structure in Marvin Sketch returns you to the dialog box that also allows you to enter an optional lower limit for similarity with the structure previously entered The example in Figure 23 searches for compounds containing the Thiophene group Similarly as in quantitative search the distribution map changes depending on the partial structure search parameters After partial structure search compounds can be selected within the distribution map and the structures displayed in the Compound List panel Figure 24 Kyoto Constella Technologies 15 Before the search CzeekD Manual After the search S347 7377 SE SOR 4 84 498864602 479 6 500 fiat porto aa a de saved Selected D Figure 22 Transformation of the distribution map after numeric search Befo
31. s required to prevent unexpected results As an example benzoic acid is set as the fragment as shown in Figure 41 Since it is a carboxylic acid based on RECAP rules the bond used should be the amide bond However in CzeekD both the amide bond and the ester bond can be used so they are presented separately In Figure 41 in case of amide bond the OH group is replaced with LA On the other hand if the ester bond is to be used the OH group is replaced with Pr If both bonds are to be used during the calculation both fragments with the appropriate labels should be included in the upload file c is fac Neon uf Table 3 Heavy metal marker pairs that allow binding O If using amide bond La O on O O If using ester bond Pr Figure 41 Selecting the heavy metal markers 5 7 Advanced calculation parameters Clicking on the Advanced button in the settings windows will open a popup window similar to that shown in Kyoto Constella Technologies Al CzeekD Manual Figure 42 The upper left hand area of the window allows the selection of the optimization algorithm PSO or GA to be used in the calculation CzeekD uses PSO by default so in this case just keep the PSO radio button checked The default number of iterations is set at 2000 but can be set higher if a more refined chemical space search is required However increasing the number of iterations will also increase the amount of time required for calculation It is recommend
32. s with attachments and are classified according to the number of attachments Fragment binding also follow the RECAP rules unless otherwise specified That is binding based only on the above combinations of heavy metal markers is allowed For example amide bonding is only allowed with fragments having La and Ce attachments In Figure 4 below combinations that are not allowed in RECAP like Lu and Am below do not result to binding and therefore to the formation of new compound structures Using the RECAP rule to govern the binding of fragments results to creation of virtual compounds with high probability of synthetic accessibility A UL SIIT SC Ay Figure 4 Example of a fragment combination not conforming to the RECAP 2 3 Particle Swarm Optimization PSO algorithm This section describes the process by which appropriate fragments are arranged in each unit The use of PSO algorithm in CzeekD allows selection of fragments for binding such that the capabilities of the scoring function is maximized Kyoto Constella Technologies CzeekD Manual PSO is an optimization technique that mimics the movement of insect swarms and fish schools that are seeking food or safety It has been found to have better performance than genetic algorithm GA Exchange of information between each particle in a swarm regarding the best solution each has found on its own leading to the determination of the overall b
33. tained and managed at Kyoto Constella Technologies and therefore it does not require any installation on the part of the user Users can access the service using the web browser installed in their PCs Internet Explorer Firefox Google Chrome Below are the system requirements to use CzeekD e OS Windows 7 8 8 1 e Screen Resolution HD 1366x768 and above Full HD 1920x1080 is recommended Browser Firefox Google Chrome Internet Explorer e Java version 6 or 7 e Flash Player Adobe Flash Player 11 and above e Network connection Broadband 1 2 Login and operational policy Users can reach the CzeekD server by accessing the URL below https denovo czeek jp czeekd Once you reach the login page fill in your login name and password then click on the Login button Please change your password particularly during your first login Refer to section 5 for details Below is our server operation policy e Our servers are structured such that our application server is our frontend and the calculation server is our backend e We dedicate one CzeekD server for every customer e Data backups are performed automatically every Sunday midnight AM e In principle we retain customer data only for a period of one year Kyoto Constella Technologies Il CzeekD Manual 2 Creation of Compounds in CzeekD The principle behind the creation of virtual compounds using CzeekD will be explained in this section 2 1 Compound creatio
Download Pdf Manuals
Related Search