This conference is all about advanced cell and tissue culture. How does in vitro fit in with in silico, read across and other methods?

Whilst in silico and read-across methods are, of course, independent from in vitro approaches, it is the practical application of alternatives within the 21^st Century Toxicology paradigm that is exciting. For instance, the extrapolation of data from in vitro assays to meaningful safety assessment for human exposure is dependent on the use of computational modelling of various types.

There are also a number of ways in which the new in vitro and HTS data can be used, thinking of the immense resource of ToxCast data as an example – can we, and how should we model these data? How can we extract usable information from these ToxCast data? Read-across is now seen as a very important technique, however the past couple of years have demonstrated that it is no longer a case of two structures looking similar, but the need to build a body of evidence – which can, in part at least, be derived easily in vitro.

We have seen the terms “New Approach Methodologies” being applied to these data and an increasing reliance on read-across to be supported in this way.

Which of these can help us develop better understanding of AOPs and do we actually need to understand or are ‘black box’ methods adequate?

My training was in biology so I believe every model should be mechanistically interpretable!

Indeed, for many years in the QSAR / in silico community there were fundamental differences between those of us who believed models should be derived from mechanisms, as opposed to those who favoured a more statistically led approach with the informatics driving the model.

There is a place for both methodologies, however an increased emphasis on AOPs (and Mode of Action) has brought the focus back to mechanisms of action. There are challenges and opportunities with AOPs, for modelling we have already turned to the Molecular Initiating Events of AOPs to help identify chemistry associated with particular events leading to liver toxicity.

Further, more ambitious, work in our group, will look at modelling of quantitative AOPs.

ACTC is a small conference with lots of opportunities for discussion. What do you hope to get out of ACTC?

I always learn from attending presentations and meetings “out of my area”.

For too long QSAR and in silico modellers did not consider the data they were modelling – often simply taking a data set and developing a model, sometimes without checking or even thinking what the endpoint was.

Coming to a meeting with an in vitro and human relevance will give me a better understanding of these areas of science and the possibilities, and pitfalls, of modelling their data. We are moving to a world that is integrating all sorts of methods, the reliance on standalone methods as replacements for animal tests has passed.

The future of areas such as safety assessment is using in silico and in vitro methods together to provide a better, more human / species relevant prediction.