chemical modelling
things simply because they don’t get the time to investigate the results from a broader perspective outside of the main aim of the study. ‘The problem is that, if the workflow is focused and geared to answering just one question, then a lot will be missed – so we have to introduce flexibility and creativity into our research,’ he says. The company’s Omics Explorer is used
by many large pharmaceutical organisations for multi-variant data and the results are presented to the user instantly, regardless of what type of update or analysis, or statistical model they select. This enables them to test different ideas in a short time and gain that ‘broader perspective’, explains Ivarsson. Running on a standard computer, Omics Explorer delivers powerful statistics and a suite of visualisations. One of the key techniques is principal component analysis (PCA), an optimal way to reduce data so that it can be presented in three dimensions.
Experimenting with models Although in silico techniques can speed up drug discovery, some labs continue to look to experimentation first as a means of solving dilemmas. Lalitha Subramanian says: ‘One of our customers was developing novel biologics and had an issue with the fact that
several batches of Active Pharmaceutical Ingredient (API) had very poor activity. They spent a couple of years trying various experiments but could not improve the potency of those biologics, so they approached us to see if our modelling and simulation technology could help.’ She notes the considerable amount of
time and money that was spent running experiments before Accelrys’ scientists came up with a unique methodology. ‘Through the analysis of 3D structures generated, our
the molecule to improve its properties, but can propose new ideas that may result in a better molecule. He points out: ‘That was a really
complicated problem, as a computer is not a chemist – it doesn’t necessarily understand what chemists can and can’t make; and the worst thing we can possibly do is give a scientist a list of potential molecules that are all meaningless. Our technology will take an initial molecule, under the assumption that it has some sort of issue you are looking to
THE PROBLEM IS THAT, IF THE WORKFLOW IS FOCUSED AND GEARED TO ANSWERING JUST ONE QUESTION, A LOT WILL BE MISSED
scientific research services team proposed modifications to improve the potency. The customer then ran experiments with the modifications we suggested and were able to fix their problem,’ she says. Having a team to offer suggestions as
to what modifications should be made to a compound, or what line of investigation should be followed is an advantage. Conscious of this, Optibrium plans to introduce a version of its software that will, Segall explains, not only help chemists to select compounds and provide feedback on how they might change
overcome, and it will generate a whole family of molecules that are related to that first one. Segall explains that this will be done in a
way that is based on the historical experience of medicinal chemistry and while it won’t guarantee that it’s easy to synthesise, it will ensure that the molecule will make sense from a medicinal chemistry perspective. ‘It will be a plausible molecule that you might want to consider,’ he says. Segal advises chemists that the models they choose to use should be robust and well validated, adding: ‘But no model is perfect.’
38 SCIENTIFIC COMPUTING WORLD
www.scientific-computing.com
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44