IT aided research and development

Bioinformatics is critical to low cost, high quality R&D. Yet the market for it seems cold. Deepali Gupta breaks down the basics of this marriage between IT and pharma research

Not long ago it was prophesised that the bioinformatics market is on the way to a meteoric rise. The influx of small and medium biotech companies clubbed with the wave of IT outsourcing to India held promise for this confluence of biotech and IT. Yet, the bioinformatics market today seems stagnant at a mere Rs 80-100 crore employing less than 200 people, reveals Dr Jayshree, Senior Research Analyst, Healthcare Practice Division, Frost & Sullivan. “A lack of profit-making ventures and very little scope for innovation have soured the bioinformatics dream,” she avers.

But does that mean these solutions have become redundant, and are immaterial to pharma and biotech companies the world over? Certainly not. It is the industry's belief that bioinformatics has now passed its hype cycle, and though its market seems to be on a slump it is now an integral part of chemical and pharma research.

What's the point?

Research is expensive, and particularly in pharma the entry barriers are high. Besides, the probability of success is so low, close to two percent, that most investors are extremely prudent while entering novel drug discovery. “Drug discovery costs have spiralled nearly 55 percent, with cost of discovering a new drug amounting to nearly $802 million,” says Jayshree. The cost of every additional day spend on drug discovery and development is estimated at $1 million.

Bioinformatics is a tool which helps facilitate and automate research to make it less error-prone. IT solutions help scientists keep a track of all the previous research to ensure the organisation is not reinventing the wheel. “Nevertheless, these solutions contribute in generation of ideas. It is not the end of the research journey. After it inspires an idea, the development still has to be taken to the lab,” clarifies Yogeshwar Rao, Head Technology Networking and Business Development Group, CSIR.

The decision to invest in a bioinformatics solution for any company looking at innovative research, is inevitable. International players already have such solutions on board, and any enterprise that hopes to compete in the search for intellectual property, this will be an imperative tool in research.

What does it do?

Under the roof of bioinformatics come multiple applications. Here are some of the functions these solutions can be used to perform:

Genome comparisons: If a particular bacterium causes an immune reaction or disease in a human being, the genomes of the bacterium can be mapped. Thereafter, it can be compared with the human genome. By isolating unique genomes in the bacteria, drug developers can identify the target of the new molecule. The advantage of the comparison is that the drug is less likely to have an adverse effect on humans because it will not target any component on the human genome.

In-silico testing: Literally, this means testing on silicon, or in a computer. The technology can be used for generating 3D models of molecules and proteins. For example, some sources of disease or infection are proteins that secrete irritants. If the usual behaviour of these proteins is known and documented, using predictive tools, molecules can be tested for whether or not they would bind with the protein and render it ineffective. Club that with a molecule modelling suite, and the largest part of trial and error in the lab can be conducted without actual lab work.

Tracking genome mutations: Certain genetic disorders are a result of early genetic mutations. If identified in time, corrective measures can be taken to prevent the manifestation of certain ailments. For example blindness in glaucoma patients. This technology is particularly useful in developing diagnostic kits.

How does it work?

The conventional image invoked by a bioinformatics solution is that of large mainframe computers and expensive equipment. However, that is no longer the case. With hardware prices at an all time low and development of intuitive software that can run on desktop PCs.

“Most of our solutions run on a Pentium IV, with 256 MB RAM and about 20 GB of free hard disk space,” explains Vibhav Garg, Functional Consultant, Mascon Life Sciences.

Most of the high-end solutions too can run on machine clusters, which means that research companies need not add expensive hardware, such as servers, to their existing infrastructure. They can use the processing power of their desktops through the network.

Data analysis tools such as the CSIR BioSuite accept sequence (for genes) or structure (for molecules) data. This is data typically synthesised in the biotech labs and input into the system in one of the seven standard formats. Numerous algorithms run searches checking for homologs and energy minimisation using conjugate methods. on the data based on the data that already exists at the backend.

For simulations and predictive tools, some amount of machine learning is required. Most modelling tools have some pre-defined learning sets, but these can be developed without much trouble. To develop a training set however the organisation needs to be data rich in the field of research.

The good and bad

The most important aspect of such a solution is the database. A well structure repertoire of past research acts like the groundwork for new research. It is based on information stored in these databases that the computer can compare and generate results for projects under development.

The good news is that a large number of such databases are readily available, both online and with service providers. If you purchase a solution from vendors, usually you get access to their databases without requiring the infrastructure or maintenance cost to host it.

The bad news is that there is no standard format for these databases. To use more than one therefore, an organisation requires it's own structure and middleware to convert or retrieve information from these databases.

One size doesn't fit all

As for data mining tools, Dr Sudhir Kulkarni, Principal Scientist, VLife Sciences Technologies explains, the domain knowledge required is high, and few pure IT vendors could explore it in adequate depth. Besides the system needs to evolve with the research. So, while it may be smart to pick a solution off the shelf, it is equally as important to identify a customisable solution and team that can alter the system along the way.

The other aspect is that if there were a one-size fits all kind of solution it would put a large number of competitors in this race for IP at a level playing field. That is perhaps why these solutions are a matter of confidentiality, and companies prefer to develop them in-house.

VLife Sciences has software packages and predictive tools to aid design for drugs as well as modelling of bio-molecules. But by and large their core focus lies on the discovery of new drugs, says Kulkarni, While marketing the bioinformatics solutions may be a side business, it is not a thrust area.

That is possibly why the market for bioinformatics has only seen sporadic success. “The 'take-off' will never happen because even if it is simple data analysis, companies need to keep it confidential and unique,” explains Kulkarni. Having common software eliminates the uniqueness.

editorial@expresspharmaonline.com