Sarah Features and Benefits
 |
Expert ICH M7 Support - Sarah predictions are accepted by regulators under the ICH M7 guideline. The M7 prediction functionality allows for simultaneous compound processing in Sarah and Derek against the mutagenicity in vitro endpoint, fulfilling both the expert rule-based and statistical-based predictions required under ICH M7. |
|
Rapid Mutagenicity Assessment - Sarah can swiftly provide single or batch predictions for the mutagenicity of query compounds. |
 |
Extensive Coverage of Chemical Space - “Out of Domain” predictions pose a significant problem when assessing impurities under the ICH M7 guideline as two negative predictions from complementary methodologies are required to conclude that an impurity is of no mutagenic concern. “Out of Domain” predictions are of little value when conducting expert review and add additional complexity to this process. Sarah’s large, high-quality training set ensures maximum coverage of chemical space thus minimising problematic “Out of Domains”. |
 |
Accurate Predictions for Proprietary Compounds - Sarah has been extensively evaluated using pharmaceutical proprietary data sets and delivers high performance (Barber et al. 2015). |
 |
Confident Predictions - The Sarah model is built upon a large, high-quality dataset that has been curated by Lhasa experts. Structures have been standardised and conflicting data points have been solidly assessed meaning you can be confident that the predictions are based on scientifically robust data. |
 |
Transparent Predictions - Predictions are clearly represented and supported by a measure of confidence, the fragments on which the predictions are based and relevant examples from the Sarah training set ordered by structural similarity to the query compound. This high level of transparency facilitates the expert review process. |
 |
User Control - Sarah has default prediction settings that have been designed to ensure the best balance between sensitivity and specificity for use under the ICH M7 guideline. However, it is recognised that users may want to change this approach for their particular need and Sarah’s intuitive prediction setup options allow you to tailor the prediction accordingly. |
 |
One Interface, Multiple Predictions - Using Sarah within the Nexus platform can give you direct access to other Lhasa in silico tools, including Meteor Nexus, Derek Nexus and Vitic Nexus, improving workflow efficiency. |
 |
Customisable Reports - Sarah incorporates a reporting framework that allows (.doc .pdf .xlsx and .sdf) file export. Report templates are fully customisable by the end user ensuring that you can provide the right information at the right time. |
 |
Auto Classification - A user editable ICH M7 classification is automatically derived from the predictions provided by Derek and Sarah and any relevant experimental information (from the Carcinogenicity Potency Database and Lhasa certified Ames data). |
Barber et al. (2015). ‘Evaluation of a statistics-based Ames mutagenicity QSAR model and interpretation of the results obtained’, Regulatory Toxicology and Pharmacology, vol. 76, April, p. 7-20. http://dx.doi.org/10.1016/j.yrtph.2015.12.006
