Chemistry in Mobile Spaces

The symposium in Denver on Chemistry in Mobile Spaces highlighted a growing number of mobile chemistry capabilities, and definitely left one with the impression that mobile chemistry research is becoming more of a reality. The symposium did not define mobile devices in advance, and so apps aimed at a variety of devices were included. In reality, there is a difference in what a person is willing to do on a smartphone versus a tablet device, and as one moves from an iPad to a small laptop or notebook, the distinction in devices becomes less clear. There are hybrid laptop/tablets which have a removable display. Plug in the display, and you have a “normal” laptop. Remove the display and you have a tablet. “Normal” is in quotation marks because at the current time, these hybrids are limited to the tablet operating systems, and so do not have the capability or flexibility of a traditional laptop computer. In that sense, we are in a transitional phase. Someday soon, we might be able to have the best of both worlds - a full strength computing environment when we need it for data entry and data processing, coupled with an ultra portable touch interface for reading and entertainment. The focus of the symposium, then, was on the current set of applications and uses for mobile computing devices. These capabilities were definitely focused on search, discovery and consumption of content. Processing and analyzing content were also discussed, but to a lesser degree, and the model emerging there is toward mobile devices interfacing with server- or cloud-based applications. This focus may in fact be a way in which cloud-based services displace local storage, so individuals with multiple devices can have a seamless experience across their devices.

Tony Williams, from RSC/ChemSpider, opened the session with an excellent introductory overview on current apps in chemistry. Included in this brief survey were: Mendeley on iPhone, Wikipedia Chemistry, Theo Gray The Elements, structure drawing for calcs (formula, mass, predictions, DB lookup), ChemJuice, ChemDoodle, ChemMobi (DiscoveryGate), name lookup for ChemSpider, structure lookup for ChemSpider (cs.m.chemspider.com), and structure searchable apps: Yield101, Reaction101 from Alex Clark.

ChemSpider Mobile was just introduced in the last few days, and has had about 250 downloads/day initially. Under development is a substructure search capability as well as reaction searching.

There are a number of chemical reaction apps, such as Named Reactions, Organic Named Reactions, ReactionFlash, Reaction101, and Yield101.

ChemSpider is looking at the appropriate capabilities within apps. When considering spectroscopy, they questioned whether processing of spectra was needed on a mobile device. The capabilities for viewing, predicting, and analyzing were considered to be useful in a mobile environment. Tony presented SpectralGame (spectralgame.com), a website that was developed by Jean-Claude Bradley, Andrew Lang and him with the purpose to train students in structure determination of organic compounds from spectra. The website does so by taking a spectrum from the ChemSpider repository and suggesting two molecules of which the user has to choose the appropriate one. Besides training, the other purpose of SpectralGame is to use crowdsourcing to curate spectra that were submitted to the ChemSpider platform. Since iOS mobile platforms don’t support Java, ChemDoodle can be used instead of the JSpecView Java plugin to view spectra.

Williams announced two apps about to be released: RSC Mobile and Chemistry World. Williams also unveiled the SciMobile Apps Wiki: scimobileapps.com, which is a site where app producers can list their app and highlight its features.

Kurt Zielenbach, from CAS, presented SciFinder Mobile. Rather than an app, this is a mobile-friendly interface to the SciFinder website that was launched in April, 2011. Because it is a web-based technology, it can be platform independent, working on multiple devices and mobile operating systems. This is important for providing broad access to users because although the iPhone and iPad are clear leaders in the device market, the operating system market is more diverse. From June, 2011, Android devices account for 54% market share to 26% for iOS. RIM, Symbian, and Windows Mobile follow with 15%, 3%, and 2% respectively, based on market research from Millennial Media.

Zielenbach noted the history of CAS in mobile devices, pioneering the first representation of a chemical structure on a BlackBerry in 2005 with CAS Mobile. He also presented some interesting statistics on smartphone usage from comscore.com. For example, in Italy and Spain, smartphone uptake is greater than 30%, leading all other countries in North America and Western Europe.

SciFinder offers the highest priority features for mobile access including explore by research topic, substance ID, author name and company name, along with ability to review Keep Me Posted alerts and saved answers. All of these were identified in a global CAS SciFinder customer survey and

implemented. SciFinder features not available on SciFinder Mobile at this time such as structure entry and structure searching were identified as lower priority by users. An important consideration for customers was that SciFinder Mobile should access personalized information from the web-based SciFinder desktop platform, including saved answers and “Keep me posted” alert results. This allows the user a uniform experience across devices.

David Evans, Elsevier Properties SA (EPSA), presented survey results showing what their users are currently using, or would like to use their mobile devices for: reading articles (75%), searching (70%), recording (60%), and analyzing data sets (50%), while only 30% are writing publications. EPSA decided to develop for Apple iOS at this time, partly due to market share, but also because the Android market was too fragmented with several different versions of the OS.

Reed Elsevier had developed a large number of iOS apps, but none specifically in this area. EPSA decided to create an app associated with Reaxys, but to focus on a single purpose and not replicate Reaxys. They presented a series of ideas to their 2010 Reaxys PhD Prize Finalists. Based upon feedback from that group, EPSA decided to create ReactionFlash, a flash card app for learning named reactions.

The development cycle for ReactionFlash addressed some of the limitations of mobile devices. Because structure drawing was not robust on the iPhone, structures were represented as pictures. However, bitmapped pictures were too large, and so the structures were converted to SVG and encrypted, so that the pictures could not be re-distributed. To enable the app to be used offline, the structures for the 200 named reactions are downloaded to the device with the app. This avoids data and roaming charges, an important requirement that the PhD students and postdocs had expressed in the development phase.

Evans also touched on the marketing aspects. Who needs the app? How large is the market? What should the price be set at? How should they promote the app? EPSA settled on two versions of the app, one a free version with 10 reactions, and the other full version with 257 reactions for $9.99. Evans pointed out that the market for chemistry apps may be too small to generate a reasonable income, and that correct pricing may be essential to achieve relevant uptake. To investigate the response to pricing, EPSA launched a short term $1.99 “back to school” offer.

Jason Wilde, Nature Publishing Group, discussed NPG’s approach to apps. He had a slightly different set of survey results, showing a significant difference in the number of apps downloaded in the last 6 months by user type. In industry, about 12 apps is the average number downloaded, with current awareness predominant. For students, approximately 8 apps is average, mainly for social networking. For faculty, about 8 apps is also the average, with learning aids becoming prominent. The time distribution of app usage was also discussed and, as expected, the time most users are using their mobile devices is early in the morning, on the way to work.

Wilde chronicled the development of NPG’s mobile products, with Nature Reader for iPhone, launching in February, 2010. A special iPad issue celebrating the 10th anniversary of the publication of the draft human genome sequence in Nature was released in July, 2010, and a special issue of Scientific American in December, 2010. The Nature Reader was updated and release for iPad in January, 2011, and a clinical medicine e-book, available via the Amazon Kindle store was released in February, 2011. Wilde noted that the Kindle had problems with scientific content, including mathematics, tables, and images. The comparison of mobile and web usage on nature.com revealed that mobile users register more often, return to the product more often, and spend more time with the product.

In thinking about the future, Wilde commented on a flipbook model, QR Codes (2-dimensional barcodes that can be photographed by a smartphone and converted to a string of text, such as a URL), and the use of HTML5/CSS3 to deliver a dynamic mobile website. Wilde also discussed some potential uses for augmented reality. For example, you position the smartphone camera over an advertisement, and see an enhanced or interactive view of the product. Or you position the camera over a structure in the PDF version of an article, and you see a 3D rendering, or links to CAS or ChemSpider.

Dan O'Brien, ACS Publications, described the ACS strategy in mobile applications. The offerings include the ACS Mobile app, launched on iPhone and iPad in spring, 2010, and on Android in March, 2011. These apps have proven to be very popular, and frequent use of the app for tracking new articles is very common. C&EN Mobile was launched for both iOS and Android in August, as a free app with a capability for in-app purchase of issues of the magazine. O’Brien discussed the pros and cons of app development, and compared these to mobile-friendly websites. The latter are becoming more functional with the advent of HTML5 and CSS3, but there is still a development and maintenance cost for developing for multiple devices. One of the issues O’Brien mentioned is the support for Unicode characters which would allow accurate rendering of the full range of characters used in scientific publications. At this point, support across browsers is not uniform, and the mapping to smartphone operating systems is not completely consistent either.

Also discussed were the technologies that ACS Publications is evaluating for future mobile offerings. The environment is changing rapidly, and a number of new technologies for addressing some of the limitations associated with smartphones are emerging. These include enhanced PDF technology, which allows for reflowing of text depending on the size of the viewing device, Wolfram’s Compound Document Format, and applications which can handle the EPUB format. Also included were a discussion of QR Codes, and the presentation was sprinkled with QR Codes to allow attendees to snap a photo of the slide, and quickly link to websites for additional information.

Theodore Gray, Wolfram Research and TouchPress, talked about his app, The Elements, which was developed using Mathematica. The app was referenced by the earlier speakers as an example of a publication which really takes advantage of the native environment of the mobile platform. Gray described his dissatisfaction with the education system and with educational materials in the US, and developed this app to provide an engaging way for students to learn about chemistry. (Note: Gray received the 2011 Grady-Stack Award from ACS for in recognition of outstanding reporting directly to the public, which materially increases the public’s knowledge and understanding of chemistry, chemical engineering, and related fields). In addition to the high quality pictures, interactivity, and advanced capabilities (3D rendering of images), the app has direct connections to the WolframAlpha computation engine which provides a host of additional properties about each element.

Steven Muskal, Eidogen-Sertanty, went through highlights of each of the apps offered by his company, many of them in conjunction with Alex Clark of Molecular Materials Informatics. Eidogen-Sertanty is committed to mobile computing, and has developed mobile apps for a variety of different uses. These apps include a mix of paid and free, and in order to provide the best capability for the mobile device, the apps make liberal use of a cloud infrastructure for data storage and calculations. Muskal noted that the growth of Amazon AWS was much larger than the growth of Amazon.com retail. Some of Eidogen-Sertanty apps include iKinase Pro, which is a paid app, and iKinase, iProtein, Mobile Reagents, Reaction101 for balancing reactions, and Yield101, which are all free. Muskal also mentioned OSRA processing of photographed structures, with name lookup.

Rich Apodaca, Metamolecular Inc., discussed HTML5, JavaScript and CSS3 as sustainable technologies to overcome the cross-platform problems that software producers are facing when developing apps for mobile devices. As a development platform he recommended PhoneGap to create cross-platform apps. This HTML5 platform allows users to author native apps with web technologies, as well as to get access to mobile APIs and app stores of the various mobile OS producers.

Metamolecular Inc. is currently working on ChemTab, software to browse chemical database files (e.g. SD files) on a tablet. ChemTab generates the structures dynamically from molfiles and uses portable JavaScript as the top layer in its software architecture. The connection to lower architecture layers such as the cheminformatics part, the structure editor, UI emulation widget, and native OS APIs is done via a bridge integration software.

Apodaca then presented ChemWriter, a structure drawing software, and ChemVector, a fast browser-based structure renderer. Both were completely written in JavaScript and require no browser plugin.

Muthukumarasamy Karthikeyan, National Chemical Laboratory in Pune, India, described an environment where hand drawn structures could be photographed, using a standalone digital camera, or the digital camera on a smartphone, and sent off to a server for conversion to a chemical structure connection table. A number of chemical OCR packages are available, and Karthikeyan presented benchmarks of the two engines CLiDE and OSRA. There are problems which are inherent in chemical OCR, and these are even more complex with hand drawn structures, where bonds may not be straight, overlapping bonds may be difficult to interpret, the drawing background may not be uniform, or color in contrast to black-white printed structures must be treated. The open source software developed at the National Chemical Laboratory of India applies some of the standard recognition algorithms used in the other chemical OCR engines, but is also able to treat the color and the background problem. In the discussion following, he acknowledged a 25-30% recognition rate for hand drawn structures. The chemical OCR is not really a mobile capability at this time, but the integration of a smartphone camera with a chemical OCR server brings the structure recognition capability to the mobile device.

In summary, there is no shortage of apps and interfaces for both research and education for smartphones and tablets, enough to justify a purchase for any chemist. How well you change your habits to incorporate the device and the apps into your daily workflow is up to you. Chances are, though, that if you aren’t already using one of these devices for mobile chemistry, you will be soon.

Martin Brändle and David Martinsen, Symposium Organizers


Five recorded presentations from this CINF symposium are available at the ACS Learning Center (http://www.softconference.com/ACSchem/)