Product Review - Reaxys
Product Review: Reaxys
Executive Summary
Reaxys is a web-based search and retrieval system for synthetic chemists, materials scientists, biochemists, medicinal chemists, geologists, and mineralogists working in academia, industry and government. It combines the data from the CrossFire Beilstein, Gmelin, and Patent Chemistry databases (which were discontinued in December 2010) into a single integrated resource providing access to extensive chemical compound and reaction information.
Overview
Reaxys is a web-based search and retrieval system for synthetic chemists, materials scientists, biochemists, medicinal chemists, geologists, and mineralogists working in academia, industry and government. As of the December 2010 update, there were 30 million reactions, 20 million substances drawn from 4 million citations in Reaxys. It combines the data from the CrossFire Beilstein, Gmelin, and Patent Chemistry databases (which were discontinued in December 2010) into a single integrated resource providing access to extensive chemical compound and reaction information. The interface is clean and much more user friendly than earlier iterations of these databases. Reaxys can be searched by structure, reaction or textual information and has the capability for the creation of very complex searches. There are many links between the different search and results screens. Results contain links to full text journal articles and patents and can be exported as pdf or Word documents as well as Excel spreadsheets.
Chemical Abstracts Service’s SciFinder is often seen as the chief competitor to Reaxys. Like Reaxys, SciFinder includes several databases. Its strength is in bibliographic, substance, and subject searching, and it contains a great deal of information on substances, reactions and syntheses. However, property data in SciFinder has never been a particularly strong point and is often difficult to locate. The recent addition of extensive property information to the substance file has corrected this to some extent; however, the fact that many of the properties are calculated rather than experimentally determined still gives Reaxys the edge in this area.
There are other compound databases available in addition to Reaxys. Some are available by subscription and others for free. Examples are ChemIDPlus and PubChem Compound from the National Library of Medicine, ChemSpider sponsored by the Royal Society of Chemistry, and the Cambridge Crystal Structure Database. The National Institute of Standards and Technology (NIST) has many databases covering special areas as well as the NIST Chemistry Webbook. These are small and narrowly focused compared to Reaxys. Some of the databases contain evaluated data like those from NIST, and some are composed of author entered information subject to no vetting. For a small customer with little or no funding available, they can be useful, but not as the sole resources for serious research.
Content Evaluation
Reaxys takes data from the previous Beilstein, Gmelin, and Patent Chemistry databases, as well as data that was not in these original three content sources, and provides a view of all related data at the substance or reaction level. The Beilstein database covered organic chemistry back to 1771. Information up to 1959 was originally obtained from the original German Beilstein Handbook of Organic Chemistry. The information came from journals, monographs, reviews, and books as well as patents. Information from 1960 to the present has been extracted from selected journals and patents. The Gmelin database covered inorganic and organometallic chemistry back to 1872. Information covering the literature from 1817-1975 was originally obtained from the German Gmelin Handbook of Inorganic and Organometallic Chemistry. Information from 1976 to the present has been extracted from selected inorganic, organometallic, physical chemistry, and physics journals. The journal coverage has recently been expanded to cover key journals outside the original subject areas such as Cell and Nature, bringing the total coverage to around 400 titles. All of the data has been experimentally validated.
Historical patent information from 1803 to 1980 is derived from the original handbooks. In-depth patent information in Reaxys draws from the USPTO and esp@cenet databases. This file includes English language patents from the US (1976- ), and World and European patent offices (1978- ) from 4 International Patent Classification code covering organic, inorganic, and organometallic chemistry. Bio-sequences are included from December 2003 onwards.
User Interface and Searchability
It is possible to search Reaxys as an anonymous user. However, a searcher can register with the system which opens up the possibility for setting preferences and creating alerts. This is done by clicking on the Register tab and selecting a username and password.
The Settings screen has dialog boxes allowing a user to change the defaults for structure editor, display and search. Once saved, these settings will be in effect at login.
Source: www.reaxys.com
Substances and Properties
A very useful feature in both the Reactions and Substances and Properties search is Generate Structure by Name. Clicking on this button opens a search box in which a user can enter a chemical name, an InChI-Key, a CAS Registry Number or a Smiles string. The most recent update has added a choice of “is”, “starts with”, “ends with”, or “contains” preceding the search box. If more than one structure is found to match the search, the user is presented with a list and asked to choose one to transfer to the query window. There the structure can be modified as needed.
The default structure editor, MarvinSketch, is a java applet and contains the basic features familiar to users of drawing programs. There are basic ring templates, common atom and bond lists, a periodical table for the choice of less common atoms, selection tools, an eraser, move and rotate tools. There are also templates for residues and more complex ring structures. For reaction searching, there is a list of choices for reaction arrows, atom mapping, and repeating groups. The pull-down menus offer a wide variety of functions and can be complicated to use. Fortunately there are key shortcuts for common tasks. It is possible to paste ChemDraw structures directly into the MarvinSketch editor. Other editors have similar features. In the online help are tips for the use of each editor and a table of conversions comparing the handling of different features. Not all editors have the same capabilities.
Once a structure is complete, the user must decide whether to conduct an As-is or a Substructure search. At any point property conditions can be added to the search using one of the Properties tabs, Form-based or Advanced. This feature can be utilized either in combination with or independent of a structure. The forms (Figure 2) cover the most common data in the areas of Identification, Physical, Spectroscopic, Bioactivity, Ecotoxicological, Natural Product, and Bibliographic data. A user can specify that data in a particular field, for example, nmr data, exists in the record, or he can enter a data point or range. This is extremely useful when only specific information is required for a broad array of substances.
Source: www.reaxys.com
A very useful feature is the index included for each search box. Clicking on the ellipses opens an index. A user can choose one or more terms to transfer to the search. This can be critical in a text search because of the large number of German terms as well as other uncontrolled text in records. It is useful for numerical searches as well in figuring out how to enter ranges.
The Advanced Properties search tab (Figure 3) offers a way to do very complex and detailed searches. A user needs to know the field tag used for the type of data sought. These can be found in two ways. First, simply type the term in the Search for Field box. Or select a term from the extensive hierarchical list. As in the Forms-based search, the user can either require the existence of data or enter terms and/or numbers. Using the index feature is very important in this search. Selected terms are transferred directly to the search box with the proper field codes and operators. Once comfortable with the tags and syntax, the user can type the search directly into the box, bypassing the menus completely.
Source: www.reaxys.com
The graphical navigation bar at the top of the results screen (Figure 4) allows one to keep track of the results analysis via breadcrumbs. The substance table display tab is the default, but a user can choose to look at the substance grid display or the citation display. The grid is useful for a quick survey of a large answer set. Results can be refined by applying one or more filters from the facets on the left of the page. A recent addition is the ability to filter by sub-structure. Other possibilities include substance-linked filters such as molecular formula or weight, physical or spectroscopic data as well as document-linked filters such as document type, author, or journal title. The tool bar allows a user to limit a search only to selected substances, to output or print results, or to change the default sort which is No. of References. Output can be a substance table, grid or citation file exported to a PDF, XML, Word, Excel, RD/File, SD/MolFile or SMILES, and can be customized as to output labels and elements.
The substance display in table format gives an overview of the available information including structure, chemical names, number of reactions and available data. Beneath the structure are buttons for displaying commercial availability which can be turned off. The Show details button opens up the complete display for a given structure. The list icon allows the user to copy the structure to the clipboard or to the query window, to use it as a substructure filter, or to display related Markush structures.
Source: www.reaxys.com
Within this window the user can develop a plan for synthesizing the selected compound. Below the window is a table of different synthetic methods with a reaction display, information about the yield and reaction conditions, and information about and links to the relevant reference.
Chosen reactions of interest are added to the plan. The user can add or remove branches or copy the plan to a new page to explore alternate pathways. Once complete, the plan can be saved or output as a PDF, Microsoft Word or RDF file. Experimental details can be included for each reaction.
Reactions
A user can start a Reaction search by using the Generate structure from name box, by transferring a structure from the Substances search window, by clicking on the Copy to Reactions Tab button, or by using the structure editor. A reaction can have both starting material(s) and product(s), or one or the other alone. There are different types of reaction arrows possible. A user can map atoms from reactant to product, specify which bonds should be made or broken, and designate reaction centers.
If only one structure is drawn, the user can specify that it must be a product, starting material, have any role, or be a reagent or catalyst. If a reaction arrow is included, this choice is not available. The reaction can be searched As-is or as a substructure on all or heteroatoms. The Form-based or Advanced Conditions tabs can be used to define search conditions. Form-based Conditions contain the most commonly used fields for a reaction search. The Advanced Conditions tab is very similar to the one for Substances. The difference is the addition of Reaction Data at the top of the hierarchy.
Display tabs for Reaction search results are Reactions and Citations. The sort default is Reaxys-Ranking but there are other choices. Filters include reaction-linked ones such as yield, solvent, reaction type or number of steps, as well as the usual document-linked ones.
The results display table gives a quick overview of reactions with key data like yield and reaction conditions along with the bibliographic information for the reference. All structures have the same features as those in a substance results display.
Text, Authors and More
This tab offers two possible search types, Form-based and Advanced. The Form-based search has a Quick search box for entering text terms using truncation and Boolean operators. There are also boxes for author, assignee, journal title, patent number and publication year. The Advanced search tab has the familiar search box and hierarchical list of field codes. Again it is advisable to use the index feature to select the proper terms to transfer to the search box. This type of search has more limited use than the other two.
A text search results display has many of the same features as the other two search types. The tabs are Citations, Reactions, Substances (Grid) and Substances (Table). This view shows the title, authors, and source information with links to full display of title and abstract as well as links to display all reactions and substances in the citation.
Product Strengths and Weaknesses
Strengths:
· The information contained in Reaxys is extensive and detailed. The component databases have been key resources in organic, inorganic, and organometallic chemistry since the late 1800s. Much of the information has been experimentally evaluated, which is a very important feature not often found in other databases. Property information has always been widely scattered and difficult to track down. Reaxys has gathered it together and presents it in an easily accessible manner.
· Earlier versions of Beilstein and Gmelin online required client software of one sort or another. This was problematic for people responsible for managing local access, as they were responsible for figuring out how to get the software out to people, how to insure all users were authenticated, how to update in a timely manner, and how to placate the Macintosh users who couldn’t use the client software on their computers.
· The Reaxys interface is clean and definitely easier to use than any of the earlier versions. The system is full of links between the three search types. Structures can easily be copied from one search to another, used as the basis for a sub-structure filter in other searches, or to start a synthesis plan. A user can easily get to reactions from the substance data display or get to the full text of the references. Often the information in the data display or the experimental details in the reactions display are so detailed the original reference isn’t needed. Results can be exported to a wide variety of tools and workflow systems for increased productivity.
· There is extensive help information, either within the search environment, or from the training center, which has lots of downloadable material. The support staff is accessible and helpful. The regular updates introduce new features and fixes for problems that have been reported. There is a newsletter and a new user forum for keeping up with developments. Elsevier has a large presence at major scientific and library meetings and representatives there are available for detailed consultations.
Weaknesses:
· The target audience for Reaxys is limited and the cost is high. In the days of CrossFire, the Beilstein and Gmelin databases could be purchased together or separately. This meant that smaller institutions or larger ones with no interest in or need for both could still buy one for a reasonable price. With Reaxys, it’s all or nothing. When the product was originally introduced, there was considerable dismay about the inability to purchase the databases separately, and the addition of the Patent Chemistry Database, which many people would not have chosen to pay for. While this model may make sense for an industrial or a large research university setting, it does make Reaxys less attainable for smaller institutions.
· Not all of the capabilities of the CrossFire system have been implemented at this time because of problems with standardization across the three component databases. This has been a problem especially for Gmelin. The ability to search for ligand molecular formula information is important for organometallic compounds. This was lost in the translation. The information has now all been processed and should be loaded in one of the next updates. Other useful are still unavailable. The reason for the lack of one-to-one data field conversions is that the data structure between the three databases had to harmonized and it wasn’t possible in the case of every field. Elsevier is working on the problem and information will be added as the issues are resolved.
· The textual information from the original Gmelin Handbook was very extensive. Unfortunately not all of it has been (or ever will be) added to Reaxys. This means the user has to have access to either the print Handbook or to someone who does, or wait for interlibrary loan to deliver the information.
· Not all the structure editors are equivalent to the full-featured drawing programs. One especially problematic area is drawing variable points of attachments in the default editor. Although the creation and search for R-groups works well, VPAs do not. Although the instructions for setting up ChemDraw as the editor are straightforward, the reviewer has had problems. However, the fact that you can cut and paste structures from ChemDraw directly into the MarvinSketch makes this less important.
Technical Requirements
Reaxys is a Web-based application using Java script and Java applets to support various structure and reaction query editors. As such it may run on all platforms which support Java 1.5 or higher and JavaScript execution. An extensive chart lists all supported operating systems, Java versions, Browsers, and structure editors. The chart also includes links to Java download and test sites as well as links to any plugins required for the alternate editors.
Service and Support
Reaxys has an extensive support system for users including:
· Online help for all functions, including tips for using all the different structure editors.
· Training Center which has numerous downloadable training materials
· Newsletter
· Free Webinars
· FAQ with many common questions
· User forum - introduced in the last update
· Phone or email help
· Feedback form
Licensing and Contracting Options
Reaxys may be licensed under an annual subscription fee on an open unlimited access model giving all users at the institution, including remote and walk-in ones, access. Specific license terms are confidential.
Pricing Model
A Reaxys subscription is based on an annual fee with unlimited usage. The price is based on the size of the user population and is confidential.
In Outsell’s Opinion
The information contained in the Beilstein and Gmelin Handbooks has always been valuable and historically, in the US at least, underused. The several online versions of first Beilstein, and later both handbooks, have not been very user friendly, both in the cumbersome nature of the search interface and the requirement for client software. Reaxys represents great strides in both these areas. The user interface is straightforward and rich with features. There is lots of help available and, best of all for those in charge of supporting and training the users, it requires only readily available web browsers and java applets. In addition, the licensing/access policies for Reaxys compare favorably with those for SciFinder, since Elsevier grants open, unlimited access. SciFinder use is limited, in the case of an academic institution, to only faculty, students, and staff at that institution. Price is based on the number of simultaneous users which results in very limited accessibility in the middle of the day unless an institution can afford a large number of users.
Although intended originally for chemists, the breadth of information in Reaxys makes this database useful to a wide range of scientists, engineers and, even, geologists. In an academic setting, it is often the resource of choice for finding properties and spectra needed by undergraduates as well as for the drug development chemists and pharmacologists looking for a perfect synthesis.
