Last week Richard Holland and myself flew to Singapore for the inaugural Bio-IT World Asia conference, organized by Cambridge Healthcare Institute (CHI). For those who missed the event and have not heard about the Bio-IT series, Bio-IT Asia brought together life sciences, pharmaceutical, informatics and IT professionals from Asia and the rest of the world. More than 250 people attended the talks presented at the stylish Marina Bay Sands convention centre in Singapore. Delegates shared advances in their research as well as how current limitations are being addressed.
Bio-IT Asia was a four day event which started with a series of pre-conference short courses, one of which was chaired by Eagle and entitled: Analysing NGS Data Using Open-Source Tools.
Eagle and a panel of four speakers, from Singapore and Japan, taught the audience how knowledge of open-source tools like eHive and Taverna can help quicker achieve NGS research goals. The panel of speakers included:
- Charles Plessy, Riken Institute – talking about DebianMed, an operating system that is particularly well suited for the requirements of biomedical research.
- Mark De Silva, National University of Singapore – talking about BioSlax, a live CD/DVD Linux distribution with a suite of bioinformatics tools.
- Toshiaki Katayama, DBCLS Japan – talking about BioRuby, a project which provides an integrated environment in bioinformatics for the Ruby language.
- Kevin Lam, National University of Singapore – who raised and attempted to answer the question of whether open source or commercial tools are better for getting research done quickly and accurately.
From left to right: Kevin Lam, Charles Plessy, Richard Holland, Mark De Silva and Toshiaki Katayama
The second day, and the first day of the main conference itself, began with a great talk by Dr. David Ho of Rockefeller University’s AIDS Research Center on the origins of HIV/AIDS and how, thanks to the advancing technologies in sequencing and genome analysis, we are now able to understand in more detail the mechanisms of the disease for which over 30 drugs have been developed but without any achieving 100% efficacy (currently the best achieves 80%).
Other keynote speakers were from BioTeam’s Chris Dagdigian, specialist in designing research computing and infrastructure, who emphasised the best practices in his Trends from the Trenches (check GigaScience’s blog about BioIT Asia), and Yaron Turpaz, VP of Informatics from AstraZeneca, who showed a great video example on how an extra pair of eyes is always needed to see all the details in research and gain maximum potential for analysing them. He mentioned the Pistoia Alliance and explained how it had been assembled for open collaboration between companies across the sector to tackle common barriers, helping decrease the time to market of a drug down to 5 years.
The rest of the conference included commercial presentations alongside research updates from academic institutes. Commercial presentations came from several well known companies in the sector pitching solutions for problems such as:
- The increasing storage demand.
- Moving data around the network.
- The management of big data, analysing it, and comparing it to a reference source.
- Comments and photos from the other talks can be found by searching Twitter for a summary of #BioIT12 tweets.
Again there were debates on whether open-software is better for the cause. The general consensus was that if you know which tools are best for the job then you can do more in less time, regardless of the licences applied to the software that you use.
In one of the talks it was said that the support for open source tools is some-what poor. That may be true for people attempting to use new tools without proper training or advice from an expert but experts are always on hand, including ones who are able to sign service level agreements that guarantee their response in times of crisis.
An easy parallel might be drawn with a driver who attempts to drive a car without ever having taken a driving lesson. That would be a struggle! It might be a bit easier to learn how to drive it himself if he chose an expensive car with extensive automation. That might work eventually, but the extra weight and maintenance required for the extra automated features would significantly increase the day-to-day running costs for the lifetime of the car, in addition having to pay extra for it in the first place.
Imagine if the driver went a different way and he hired a skilled chauffeur to drive a less expensive non-automated car. The chauffeur would cost money it is true, but he would use it more efficiently as he would know in great detail how the machine worked, and he would also know the best short-cuts to the required destinations. The running costs would be lower due to having less of those expensive automated gadgets on board, and best of all the chauffeur could eventually teach by example and hand over control of the car to the owner, who would then be able to drive it better than he could ever have achieved on his own with the automated model.
It is not a huge leap of the imagination to see how this could be reflected in bioinformatics. Expensive commercial software packages that come with lots of automated bells and whistles are great for getting a quick start, but they; hide the advanced options, are not flexible, and come with a high on-going cost of ownership through licence fees. Open-source software, installed and configured with the use of a professional and experienced bioinformatician ‘chauffeur’, may initially be more expensive to set up, even though the software itself is free, but in the longer term it will offer larger savings, more flexibility, and a greater ability to get your research done exactly the way that you need it.