Meeting in the middle?

Share this on social media:

Steve Maginn looks at the two ends of the chemistry software industry

It is difficult, if not impossible, for any person to give an objective view of a 10-year period in anything. It is not only things that change - people change too. Ten years ago, I was a 20-something working as a materials science support scientist for a company called Molecular Simulations Inc. (MSI) in Cambridge, UK. Now I am a 30-something director of scientific services for Chemical Computing Group (CCG), ironically based in the same building in Cambridge where MSI was located when I first joined them in 1993! My job (and location!) has changed several times in the interim, but what has changed in the industry in which I work?

By 'the industry', I mean computational chemistry. Ten years ago, there were many players at the simulations end of the industry. Some were larger than others, but there were many niche companies providing targeted solutions with particular applications, both in the life and materials science areas. MSI had recently merged with Polygen, and was active in both life and materials, as was its main competitor, Biosym. There was keen rivalry between those two companies, assuaged only when they themselves merged not long after. Tripos was a major force in life sciences, and its name alone among the large companies survives to this day. Others, such as BioCad, Synopsys, Oxford Molecular, Chemical Design etc. have disappeared, most gathered into the rolling stone that is now known as Accelrys.

It is not completely true to say, though, that the industry has rationalised. New forces have emerged. Schrodinger Inc. has grown from a niche supplier into a more general provider of software. My own company, Chemical Computing Group, did not even exist 10 years ago, but has grown rapidly to become one of the top suppliers. Other names like SciTegic and OpenEye also provide some general applications, whereas new niche suppliers such as CCDC or InforSense provide software in specialist areas or frameworks for a computational chemistry infrastructure.

So, 10 years ago there were a few big industry names and many smaller companies. Today, we find ourselves in the same situation.

Scientifically and technologically, there have been enormous advances. The range and quality of computational applications has grown rapidly, aided by the available computing power, greater volumes of and quality of experimental data (comparison of the crystallographic databases, PDB and CSD, now and then, provides a perfect example of this), and simple scientific progress. It is a well-worn cliché that each of us has much more power in our desktop and laptop machines now than was available in the most expensive, top of the range workstations 10 years ago. But looking beyond all that, has there been a change in how computational chemistry simulation software is used?

The Holy Grail 10 years ago was to make computational chemistry simulation software available to all chemists on their desktop. Has that been achieved?

'The industry' I have been referring to so far is the one supplying simulation software to computational chemists and molecular modelling groups. The industry supplying software to bench chemists was somewhat different 10 years ago, and remains so now. The range of applications has increased there too. Suppliers, such as CambridgeSoft and others, have grown from providing software for sketching 2D chemical diagrams out into suppliers of electronic lab notebooks, and relevant calculations of spectra, etc. In many cases, they have successfully joined up with suppliers of databases and database infrastructures, such as MDL, to provide informatics services.

But, despite some collaborations and company acquisitions, we haven't quite met in the middle yet. We can still talk about two 'ends' of the industry - there is not yet a continuum. There may be many reasons why this is the case, but this is not the forum in which to examine them.

In some companies, their computational chemists provide software infrastructure to bench chemists. These are custom-made, using pieces supplied by different vendors or developed in-house. These pieces include applications, data sources, file format parsers, workflow definers, etc. We still haven't seen one company providing a total computational infrastructure that satisfies the needs of both 'ends' of the computational chemistry market.

Perhaps this is a good thing! We are told competition is healthy, and competition still exists and is as strong as it was 10 years ago, despite all the mergers and acquisitions that have occurred. And as then, there are also plenty of opportunities (but seemingly little time!) for collaboration. Certainly the two 'ends' of the industry have grown closer and perhaps that 'continuum' I mentioned, that meeting in the middle, may be achieved within the next 10 years.

Steve Maginn is Director of Scientific Services for Chemical Computing Group