DECEMBER 2014 REAXYS R201 ADVANCED STRUCTURE SEARCHING

Transcription

1 DECEMBER 2014 REAXYS R201 ADVANCED STRUCTURE SEARCHING 1

2 NOTES ON REAXYS R201 THIS PRESENTATION COMMENTS AND SUMMARY Outlines how to: a. Perform Substructure and Similarity searches b. Use the functions on the right of the Reaxys structure query screen c. Use the functions built into the Structure Drawing Editor (MarvinSketch is used as the example) For an outline of other presentations in this series, go here While entries of substances in databases closely follow the normal representations used by scientists (e.g., as valence bond structures), nevertheless there may be important differences The best way to learn about these differences is to understand representations of substances in answers retrieved It is important to understand structure search defaults in Reaxys Other structure search systems may have different defaults, and thus results may not be comparable Some functions are implemented in the Reaxys search interface; others are implemented through the Structure Drawing Editor(s) It usually is better to keep structure queries simple Research scientists rarely need to build complicated queries Also remember that Reaxys has numerous postprocessing tools (Filter by: and Analysis View) to help narrow answers to those of most interest Other presentations that relate to how to find substances include: Reaxys R103 - how to search for specific substances Includes how to search by structure As drawn Reaxys R202 - how to search for groups of substances through (part)name or (part)formula searches 2

3 CONTENTS Introduction to Substructure Searching Revision of search: As drawn o Markush structure search Substructure Search o General o Ring closure o Narrowing answers o o Fragments Tautomers Functions through the Structure Editors o Reaxys generics o Generic atoms o o Editing atoms Editing bonds Similarity Supplementary Slides 3

4 Where possible, substances are entered in Reaxys as structures based on valence bond rules, e.g., aspirin However, there are many cases where simple valence bond structures cannot be drawn for substances, and these include (among others): Inorganic compounds Salts Aromatic structures/resonance π-complexes Donor(σ-) complexes Polymers There are still further cases where substances may exist as mixtures and these include (among others): Tautomers Cyclic/open chain forms (e.g., carbohydrates) Alloys Ceramics Additional issues include: Entry of substances in databases Whether the author(s) fully described substances, e.g., where stereochemistry is possible has absolute stereochemistry been assigned or is stereochemistry unspecified? Whether different physical or crystalline states are separately registered The registration of reactive intermediates INDEXING OF SUBSTANCES IN DATABASES SEARCH TYPES AND DEFAULTS Search types: As drawn (default) retrieves substances with the same atoms/bonds as in the query structure, including isotopes, stereoisomers, and salts and mixtures (where one of the components contains the atoms/bonds in the query structure) Variable structure features may be entered through the structure editor(s), e.g. R-Groups, Atom Lists, Substitution counts, and locked atoms Generic atoms (e.g., X), generic groups (e.g., CYC), and variable bonds are allowed in queries Substructure retrieves additional atoms/groups (where allowed) on the query structure on heteroatom - only allows substutions at heteroatoms on all atoms - most general substructure query Similarity retrieves substances that match the query plus substances that are positional isomers or present various levels of similarity to the query structure A table with various levels of similarities is first presented Defaults include: Structure search types and defaults Additional and fused rings allowed Chain bonds drawn in queries match ring/chain bonds in structures All allowed stereochemistries, valencies, isotopes, charges at atoms and so forth Defaults may be overridden through functions in Structure Query Editor(s) A basic understanding of substance indexing and of search types and defaults helps structure search efficiency and effectiveness 4

5 GENERAL APPROACH TO SUBSTRUCTURE SEARCHES The evolution of structure searches The ability to search for structures became available through digital storage and search techniques in the 1980s and opened enormous opportunities for scientists However, data storage and search processes were expensive and substantial costs were associated with each search To counter this, users (particularly information professionals) requested great flexibility in search queries to minimize the number of searches, so that costs could be minimized Accordingly, searches tended to become very complex, and lots of variables were built into queries including the ability to nest R-groups into different R-groups and to require different positions of substitution and different numbers of repeating atoms or groups However most of this has changed, and structure searches are available through subscription pricing rather than individual search costs Further, substances databases have increased ~20-fold in the last 30 years, and generic searches with lots of variables may lead to very large answer sets So the structure searcher may no longer need to build complicated queries Indeed the focus usually should be on designing separate searches with the emphasis on simplicity Options to obtain more precise answers include: Specifying more atoms in queries Stopping further substitution at specific positions Restricting the formation of rings Defining attachments more precisely (e.g., using fewer generic groups) Defining bonds more precisely Excluding unwanted classes of substances (e.g., salts, isotopes and so forth) Options to broaden answers include: Precision vs Comprehension Searching default values (e.g., any valency, any rings/chains, any charges and so forth generally defaults allow for greatest variation) Specifying fewer atoms in queries Leaving more positions open for substitution Using generic groups (e.g., X for halogens) Trying Similarity search Also keep in mind that often the information/properties on substances are of most interest, so a search that may give thousands of substances may contain only a few substances that have the information of interest (e.g., of a particular spectral or physical property or biological activity) The ability, in Reaxys, to search for substructure and for information/property at the same time helps substantially to find precise answers It helps if research scientists keep substructure queries simple It usually is better to build a couple of simple queries rather than to try to build everything into a single complex structure 5

6 Different substance databases in Reaxys give different information A structure search for ritalin As drawn gives: > 70 substances in the Reaxys substance database > 90 substances in PubChem 7 substances in emolecules Answers include many salts, stereoisomers and isotopic substances, and information on these substances in the different databases is different In many cases more than one registered substance may need to be explored in order to retrieve all the information required Answers may be combined (go to the History Tab and follow instructions) to get a single list of unique substances Note: Answers are sorted by default by: No of references STRUCTURE SEARCH: AS DRAWN REVISION OF REAXYS R103 6

7 Generic structures in patents SEARCHING FOR MARKUSH STRUCTURES At times it may be necessary for a research scientist to be aware of patents that cover generic structures relating to substances of interest, and in such cases the structure search may be expanded to Include related Markush structures In the case of ritalin, an additional 6 generic structures are retrieved These structures are shown below, while two of the citations (and part of the details associated with one citation) are shown on the next slide Discussion of Markush structures is outside the scope of this presentation; the information is shown here merely to give research scientists an idea of type of information retrieved through the Include related Markush function 7

8 PATENT CITATIONS AND DATA There is very extensive coverage in Reaxys of information on substances from patents (US, WIPO, EP) Records from patents contain considerable detailed information (click to see some of the information from answer #1) 8

9 SUBSTRUCTURE SEARCH Before you undertake a substructure search, consider where you would like additional substitution For example, in the case of the ritalin query (shown on right), a substructure search would ignore the hydrogens in the query structure, but you would retrieve only esters (To retrieve acids as well as esters, you would remove the carbon on the oxygen) If you search the query structure shown, all the substructure defaults apply Note the numbers of answers in the different databases Some of the answers are shown Note that some contain additional rings The icon indicates that the substance is commercially available The icon indicates that the substance appears in more than one database, e.g., in the Reaxys substance database and in PubChem 9

10 RING CLOSURE SPECIFYING THE NUMBER OF RINGS The total number of ring closures to be found in retrieved structures is defined as the smallest number of ring bonds that must be broken in order to convert the structure into an acyclic structure When No ring closures is checked, Reaxys looks at the total number of rings in the query and retrieves only substances with the same number of rings (in the case of the ritalin search only substances with 2 rings are retrieved) RING CLOSURE = 3 RING CLOSURE = 3 RING CLOSURE = 4 RING CLOSURE = 5 Effectively this blocks further ring fusion and stops rings formed from chain bonds in the query The # of Ring Closures required may also be specified, e.g., 3, 3-5 and so forth Some representative answers are shown on the right 10

11 OPTIONS TO NARROW ANSWERS FROM SUBSTRUCTURE SEARCHES THROUGH THE REAXYS SEARCH ENGINE Sometimes it is better to do a more general structure search first to understand the range of possible substances So having first done a more general search for esters with the ritalin structure you may browse through the Reaxys and PubChem answers and then narrow them if needed You may narrow answers through the Reaxys search engine* by: Eliminating some of the classes of substances through choices on the right of the structure query screen Searching a more specific structure under Filter by: (Substructure) Browsing through options under Analysis View Adding additional search fields (e.g., property fields) Remember that there are many subheadings for options under Analysis View including Compounds (classified) by: Functional Group Classification Richter Classification Ring Classification These sub-headings have further sub-classes which are seen by clicking the icon next to the option of interest OPTIONS TO NARROW ANSWERS * Later in this presentation, the various options to narrow answers through the Structure Editors are discussed 11

12 FRAGMENTS KEEP FRAGMENTS: SEPARATE/TOGETHER Separate fragments may be drawn in the structure editor, and the fragments may be searched: Keep Fragments Separate: Query and Sample Answers Keep Fragments Together: Query and Sample Answers Separate answers contain the fragments in different components of a registered substance This option is most useful for finding salts and mixtures where specific structures are required in the separate components Together answers contain the fragments in the same structure This option is most useful if two groups of atoms are required in answers, but where specific connections between the groups are not required 12

13 TAUTOMERS By default, Reaxys does not search for alternative tautomers when the query structure is a single tautomeric form However, when is checked then tautomeric forms are additionally retrieved For example, a classic tautomer question is: Is it 2-hydroxypyridine or 2- pyridone? A substructure search on query 2- hydroxypyridine gives >200,000 substances, but when the same structure is searched and Include tautomers is checked then >400,000 answers are retrieved Accordingly, when structure queries may exist in tautomeric forms and this applies particularly to many heterocyclic substances of interest to medicinal chemists (and others) then the search should include tautomers Once initial answers have been obtained and browsed, then answers may be narrowed to restrict answers to those of most interest QUERY QUERY (Include tautomers) SOME ANSWERS SOME ADDITIONAL ANSWERS 13

14 FUNCTIONS AVAILABLE THROUGH THE STRUCTURE EDITOR(S) Structure search functions implemented through the structure editor(s) REAXYS GENERICS Several structure editors may be used with Reaxys, and for a general description go to /kw/structure%20editor Each structure editor has a slightly different interface For detailed information on structure drawing with MarvinSketch, go to the ChemAxon site This presentation, as with others in this series, gives examples using the MarvinSketch structure editor Once the structure is drawn, click and the structure plus any additional query requirements specified within the structure editor are transferred to the Reaxys search engine The functions available through include the Reaxys Generics (this slide) and Generic Atoms (next slide) Reaxys Generics allow certain types of rings and chains to be searched; those ending in H allow the substances with the generic group or H (hydrogen) to be retrieved Note that only one bond may be connected to any of these generic groups MARVINSKETCH STRUCTURE DRAWING SCREEN REAXYS GENERICS 14

15 FUNCTIONS AVAILABLE THROUGH THE STRUCTURE EDITOR(S) GENERIC ATOMS The Generic Atoms drop-down menu allows generic atoms to be inserted into queries The 4 groups at the bottom of the list allow any groups to be inserted, although unlike the Reaxys Generics (previous slide) these groups allow up to two bonds to be attached That is, query (A) is acceptable (and gives several hundred answers) but query (B) is not (and gives no answers) Groups G* and GH* additionally allow rings to be connected through the G group drawn (A) Acceptable two connections allowed on G groups (B) Not acceptable only one connection allowed on CYC 15

16 FUNCTIONS AVAILABLE THROUGH THE STRUCTURE EDITOR(S) GENERIC ATOMS The query (A) (previous slide) gives around 400 substances with fused, and additional, rings Note: all searches in this presentation are intended only to give initial ideas but of course initial answers may be narrowed with additional tools For example, if the intention (query (A)) was to find only phenyl rings connected, by any group, to anthracene rings then one option may be to specify the number of Ring Closures (to 4 ring closures) 16

17 FUNCTIONS AVAILABLE THROUGH THE STRUCTURE EDITOR(S) Note: the features described here focus on MarvinSketch; other editors may look different. EDITING ATOMS The two ways to edit atoms are: a. Click on Object at the top of the MarvinSketch editor, or b. Right-click on an atom in the structure Both operations display that further links to the menu (left below) In turn this menu links to further choices that are atom specific (e.g., if you right-click on a N in a structure then the choices relate to N atoms) Note the defaults these allow for all variations; when you Edit Atom you are requiring defaults to be overridden and thus are requiring more specific options For comments on and examples of the use of some of these options, see next slide Further links from Edit Atom Edit Atom => Stereo Edit Atom => Charge Edit Atom => Valence Edit Atom => Radical Edit Atom => Isotope Edit Atom => Subs Count Edit Atom => Link Node Edit Atom => R-group 17

18 FUNCTIONS AVAILABLE THROUGH THE STRUCTURE EDITOR(S) Note: the features described here focus on MarvinSketch, other editors may look different. EDITING ATOMS Option Comments (Use the defaults unless you want to apply specific values) Stereo Charge Valence Radical Stereo (Reaction) used for reaction queries Stereo (Enhanced) used to specify stereo-centers (OR /AND are equivalent and numbers are ignored: they find compounds with relative stereochemistry) Specify charge only if essential, e.g., for reactive intermediates or when specific charges on metals in coordination compounds are required Relates to the number of bonds attached to the atom, e.g., the sulfur in a sulfoxide R S(=O)-R has a valence 4. Example here. Helpful for searching radical intermediates. Example here. Isotope Substitution count Link Node R-group Use for specifying specific isotopes. Example here. Relates to the number of attached atoms, e.g., the sulfur in a sulfoxide R S(=O)-R has a substitution count 3. You must use As drawn (Substructure search ignores substitution count) but As drawn will block substitution at other positions (unless they are opened, e.g., with substitution counts) Use for specifying varying numbers of repeating atoms. Example here. Use to specify your own variations in atoms. The workflow depends on the structure editor used. Go to Help (in Reaxys), enter r-group, and choose structure editor of your choice, e.g., for a description of the basic functions for MarvinSketch go to Note: Examples of other options through MarvinSketch are shown here 18

19 FUNCTIONS AVAILABLE THROUGH THE STRUCTURE EDITOR(S) Note: the features described here focus on MarvinSketch, other editors may look different. EDITING BONDS The two ways to edit bonds are: a. Click on Object at the top of the MarvinSketch editor, or b. Right-click on a bond in the structure Both operations display that further links to the menu (left below) In turn this menu links to further choices that are bond specific Note the defaults these allow for all variations; when you Edit Bond you are requiring defaults to be overridden and thus are requiring more specific options For comments on, and examples of, the use of some of these options, see next slide Further Links from Edit Bond Edit Bond => Type Edit Bond => Topology Edit Bond => Reacting Center 19

20 FUNCTIONS AVAILABLE THROUGH THE STRUCTURE EDITOR(S) Note: the features described here focus on MarvinSketch. Other editors may look different. EDITING BONDS Option Comments (Use the defaults unless you want to apply specific values) Type Topology Reacting Center Bond types follow the usual valence bond conventions for single, double and triple bonds Aromatic bonds are used for those substances with 4n+2 π-electrons (n= no. of rings) in cyclic systems with even numbers of atoms in the ring For a description of Bond Type: Coordinate, go here None (the default) allows a chain bond in a query to be in a ring or chain bond in an answer In Ring requires the bond to be in a ring only In Chain requires the bond to be in a chain only Example here The options relate to drawing structures for searching reactions, and are discussed in Reaxys R203 20

21 SIMILARITY The similarity of structures depends very much on the eye of the beholder So the key points with similarity structure searches are: To be aware that Reaxys contains a number of algorithms that look at similarities in structures To browse through substances in the table that Reaxys displays initially For further discussion of Similarity search see: answers/detail/a_id/3894/p/9 459/kw/similarity Some substances retrieved (Medium) Note that Similarity search for reaction diagrams work differently: it determines the actual transformation involved and then looks for reactions involving more atom spheres around the reaction center (See Reaxys R203) 21

22 SUPPLEMENTARY SLIDES 22

23 FURTHER PRESENTATIONS IN THIS SERIES REAXYS R101 REAXYS R102 REAXYS R103 REAXYS R104 REAXYS R105 A QUICK GUIDE An overview of the functions of Reaxys 2014 TIPS FOR LITERATURE SEARCHING A quick guide to searching literature BASIC SUBSTANCE QUERIES A quick guide to searching for specific substances by name, formula and structure BASIC REACTION QUERIES A quick guide to searching for chemical reactions SEARCHING PROPERTIES A quick guide to searching for properties REAXYS R201 REAXYS R202 REAXYS R203 ADVANCED STRUCTURE SEARCHING ADVANCED NAME AND FORMULA SEARCHING ADVANCED REACTION QUERIES Searching by substructure Searching for Substances by Names and Formulas Additional ways to search for reactions Back to referring Slide 23

24 VALENCE Valence relates to the number of bonds attached to the atom A substructure search on alliin (below) gives sulfoxides and sulfones QUERY Only sulfoxides are retrieved if you specify valence = 4 at the sulfur (right-click the S, then Edit Atom=>Valence (4)) SAMPLE ANSWERS Higher valencies may occur for most metals, halogens, and for N (e.g., in NO 2 and N-oxides), P, S, Se and so forth, where it usually is necessary to block higher valencies when only lower valencies are of interest Back to referring slide 24

25 RADICALS The search specifically for radicals i. Draw the structure ii. Right-click on atom where you want the radical iii.edit Atom => Radical => (Monovalent) iv. Transfer Query to Reaxys You may search for all radicals of this part-structure however, you may always perform a more specific property search at the same time, e.g., for substructures that also have ESR Spectroscopy data This search gives several hundred answers; part of three of the answers are shown on the right Back to referring slide 25

26 ISOTOPES To search for a specific isotope, right-click on the atom then Edit Atom => Isotope => <value> The table of isotopes is specific to the atom selected For example, if you have drawn U (uranium) in your query and wish to find only 235 U- containing substances, then right-click on the U => Edit Atom => Isotope (U-235) Back to referring slide 26

27 LINK NODES (REPEATING ATOMS/GROUPS) Use Link Nodes when you wish atoms or groups in a query to repeat a certain number of times So if the requirement was to find 1,3-diaminocycloalkanes (perhaps with special interest when the two amino groups are on the same side of the ring, i.e., cis-), one approach may be to the query then draw Such a query allows one of the carbons to repeat between 1 and 6 times, and answers would be 1,3-diaminocyclobutanes through to 1,3-diaminocyclononanes To draw the query is easy. Simply: i. Draw a 4-membered ring and attach two nitrogens to the ring ii. Right-click the node to be repeated (one of the other two carbons) iii. Edit Atom => Link Node iv. Choose L1-6 (see Slide 17 if you wish to see the Link Node screen again) v. Choose a Substructure search on all atoms Easy! However, the fun starts when you see some of the answers. You retrieve around 80,000 answers that contain all types of substituents and all types of rings Perhaps, at this stage, recall the comment on slide 5: the structure searcher may not need to build complicated queries; indeed the focus usually should be on designing separate searches with the emphasis on simplicity (see end of next slide for a simple approach) Two of the answers 27

28 LINK NODES (REPEATING ATOMS/GROUPS) However, let s continue and a next step may be to restrict the substructure search to allow substituents only on heteroatoms All the ring carbons now have hydrogen attachments only, but you still get well > 4,000 answers You now may limit answers to 1 Ring Closure: 1, in which case you obtain diamine cycloalkanes of interest including (1), (2), and (3) The use of Link Nodes certainly worked, but general structures may give lots of answers that may need to be narrowed (1) (2) (3) Another approach may be to build, and search, the cis-1,3-diaminocycloalkanes separately For example the cyclobutane query (Substructure on heteroatoms) quickly affords a number of substances of which (4), (5) and (6) are examples (4) (5) (6) Back to referring slide 28

29 A description, with worked examples, of the basic functions for MarvinSketch may be found at: The User s Guide may be viewed in full at: Examples of the ways to specify Position Variation and Repeating Units are shown below MARVIN SKETCH Additionally, through the basic functions link (above) you will find details of the steps required to set up R-groups Back to referring slide 29

30 SPECIFYING BONDS TOPOLOGY The most common use of Edit bond => Topology is to exclude the formation of rings (in this case you right-click the bond, then Edit bond => Topology => In chain) However, at times you may wish to force ring closure between two atoms but you do not want to specify the nature of the atoms in the connection In this case you could right-click on the bond(s) of interest, then Edit bond => Topology => In ring (left query below) You could achieve a similar result by connecting the two atoms with a G group (right query below) Some answers are shown below. Perhaps work out which structures would be retrieved in either one (or both) of the queries above. The point is that bond topology, like so many of the functions discussed in this presentation, opens interesting search options! Back to referring slide 30

31 COORDINATION COMPOUNDS To retrieve substances with coordination bonds between metal atoms and heteroatoms (such as O, N, P, halogens and so forth) it is sufficient to draw single bonds between the atoms involved The coordinate bond may be used for π-complexes, and the workflow (for MarvinSketch) is shown below However, coordination compounds may be represented in many ways and you may wish to contact the Reaxys Help Desk if you have specific problems that need to be solved Back to referring slide 31