In silico toxicology project 2021 review

The in silico toxicology project objective is to support the acceptance and implementation of in silico toxicology through working groups and publications covering: (1) protocols, (2) position papers, (3) case studies, (4) fit-for-purpose evaluations, and (5) structure-activity relationships.

As we look back at 2021, it’s been another great year of progress. So far this year, 8 papers from this effort and related initiatives have been accepted for publication, with many of these papers appearing in a special issue of the journal of Computational Toxicology.

  • Bassan A. et al., (2021), In silico approaches in organ toxicity hazard assessment: current status and future needs in predicting liver toxicity, Computational Toxicology, 20, 100187 https://doi.org/10.1016/j.comtox.2021.100187
  • Bassan A. et al., (2021), In silico approaches in organ toxicity hazard assessment: current status and future needs for predicting heart, kidney and lung toxicities, Computational Toxicology, 20, 100188 https://doi.org/10.1016/j.comtox.2021.100188
  • Bercu J. et al., (2021), A cross-industry collaboration to assess if acute oral toxicity (Q)SAR models are fit-for-purpose for GHS classification and labelling, Regulatory Toxicology and Pharmacology, Volume 120, March 2021, 104843 https://doi.org/10.1016/j.yrtph.2020.104843
  • Cross K.P. and Ponting D.J., (2021), Developing Structure-Activity Relationships for N-Nitrosamine Activity, Computational Toxicology, 20, 100186 https://doi.org/10.1016/j.comtox.2021.100186
  • Johnson C. et al., (2021), Evaluating Confidence in Toxicity Assessments Based on Experimental Data and In Silico Predictions, Computational Toxicology, 100204 https://doi.org/10.1016/j.comtox.2021.100204
  • Myatt G.J. et al., (2021), Implementation of in silico toxicology protocols within a visual and interactive hazard assessment platform, Computational Toxicology, 100201 https://doi.org/10.1016/j.comtox.2021.100201
  • Myatt G.J. et al., (2021), Increasing the acceptance of in silico toxicology through development of protocols and position papers, Computational Toxicology
  • Tice R. et al., (2021), In silico approaches in carcinogenicity hazard assessment: current status and future needs, Computational Toxicology, 20, 100191 https://doi.org/10.1016/j.comtox.2021.100191

There are a number of on-going working group activities making progress in the areas of N-nitrosamine SAR potency, carcinogenicity, endocrine activity, biomolecule reactivity, acute toxicity, and neurotoxicity. And over the coming year, we are looking to establish a series of new working groups.

Thanks to everyone from over 70 different organizations who have participated in these working groups.

If you are interested in discussing the project in more detail, please send me an email (Glenn Myatt; glenn.myatt@instem.com).

Wishing everyone a Happy Holiday! We’ll be back again in the New Year!

New posters and papers

We recently published a blog describing four new publications.1 Since this post, we have received news that 2 more papers, submitted earlier this year to the Journal of Computational Toxicology, have been published:

  1. Evaluating Confidence in Toxicity Assessments Based on Experimental Data and In Silico Predictions2

Reliability, relevance, and confidence are defined within the context of in silico analyses. Practical examples show how to assess model results by applying the concepts, including how in silico and experimental data are combined in weight of evidence assessments.

  1. Implementation of in silico toxicology protocols within a visual and interactive hazard assessment platform3

This paper outlines an interactive and visual hazard assessment computational platform and describes how in silico models are developed and incorporated based on published in silico toxicology protocols. The use of the platform is illustrated with four case studies.

These publications2-3 will appear in a special issue of the journal on the in silico toxicology protocol initiative.

We would like to thank all co-authors for their contribution to these manuscripts.

We also presented three posters at the recent American College of Toxicology meeting:

  1. Assessing Abuse Liability using Read-across and Structural Alerts
  2. Development of a Structure-Activity Relationship Profiler to Predict Mechanism-Based Inhibition of a Metabolite on CYP Enzymes
  3. Using Metabolically Similar Analogs in Read-Across to Establish Dialkyl-N-Nitrosamine Potency

If you would like a copy of these posters or would like to discuss the findings in these manuscripts, please contact me (Glenn Myatt, glenn.myatt@instem.com)

References

  1. https://insilicoinsider.blog/2021/09/23/four-papers-accepted-for-publication/
  2. Johnson C. et al., (2021), Evaluating Confidence in Toxicity Assessments Based on Experimental Data and In Silico Predictions, Computational Toxicology, 100204 https://doi.org/10.1016/j.comtox.2021.100204
  3. Myatt G.J. et al., (2021), Implementation of in silico toxicology protocols within a visual and interactive hazard assessment platform, Computational Toxicology, 100201 https://doi.org/10.1016/j.comtox.2021.100201

Leadscope database update process and statistics

The development of databases is fundamentally important to in silico toxicology. As part of our yearly release, we update Leadscope’s databases to capture either new or existing information in the public domain or data received from our collaborative partners. The process involves careful curation of chemical structures by normalization, registering the structures and resolving any conflicting registry records. The data is associated with these structures, where applicable through definition of source to ToxML mappings to build internal content. After a series of quality checks, the database content is released. There are currently over 200,000 chemical structures and almost 600,000 studies in our databases.

This process is completed for several endpoints, including those listed in the table below. For a particular endpoint, a chemical may have multiple studies conducted and although the overall call presents a resolution of the replicated studies, the individual study calls are preserved to facilitate a transparent analysis of the data. Such databases could be used to support analog searching for either read-across or to facilitate an expert review. Additionally, structures are also converted to a structure activity relationship (SAR) form, allowing these databases to be used for the development of statistical and expert alert models.  The table provides our latest database statistics. These statistics are evolving as we capture additional information that becomes available.

a. All Leadscope databases: SAR Genetic Toxicity, SAR Carcinogenicity, Leadscope toxicity databases, Miscellaneous databases (referred to as level I databases) – Drugs Chronic/Subchronic Database, Drugs Genetox Database, Drugs Repro-Developmental Database, Food Safety Acute Toxicity Database, Food Safety Chronic/Subchronic Database, Food Safety Genetox Database, Food Safety Repro-Developmental Database

If you would like any further information please get in touch at candice.johnson@instem.com.

Availability of Original CPDB Data in Original Website Format

The Carcinogenicity Potency Data Bank or CPDB has long been an authoritative and trusted source of in vivo carcinogenicity data used in toxicological assessments worldwide. Unfortunately, for over 10 years the database has not been updated and all the original formatted and unaltered content is no longer publicly available.

For many years, Leadscope has included all of the CPDB data exactly as recorded in the original database within the Leadscope databases, alongside other trusted sources of toxicological information such FDA dossiers, ECHA, and NTP data. This information is extensively used to support hazard and risk assessments, including in silico analyses. Despite availability within our products, we continually get requests to provide access to the original content in its original display format and to provide availability from a website.

We are happy to announce that in the latest software release, we have linked all chemicals with CPDB references directly to the original content in the original format. We have also made a website of CPDB content publicly, and freely available and provide links to these web pages from within Leadscope products.

If you would like to discuss in more detail, please get in touch (Glenn Myatt; glenn.myatt@instem.com).

New SD file submission recommendation from the US Food and Drug Administration

Last month, we reviewed some of our Research Collaboration Agreement (RCA) activities with the US Food and Drug Administration (FDA).1 This included a joint project to develop an internal FDA/CDER system (based on the Leadscope technology) to enter, validate, and register chemicals linked to internal (Q)SAR consultation reports. This information, along with access to toxicity databases, is made available internally to FDA/CDER reviewers through a simple web-based user interface. As part of the current FDA/CDER workflow, chemical structures are manually re-drawn from the PDF submissions before being validated, registered, and stored in the database.

The US FDA has recently made several presentations describing the benefits of avoiding this manual re-drawing of chemical structures by requesting Drug Master File (DMF) Holders to submit an electronic record of these chemicals.2,3 Such chemicals would include starting materials, impurities, intermediates, by-products, and degradants. Having access to an electronic representation of the chemical structures would avoid the time-consuming process of isolating and re-drawing the chemicals, which is also prone to errors. This in turn enables a more efficient use of resources.

It is now possible to submit structures in a computer-readable format, as an SD file, since it has been added as an acceptable file format as part of the Electronic Common Technical Document (eCTD)5. The SD file format is a commonly used electronic representation of a chemical structure and associated tabular data.4 This information has recently been posted on the FDA’s website.5

For DMF Holders to implement a process, a trusted technology is needed for collecting, formatting, annotating, and validating their SD files for submission. To address this need, we have developed a solution for creating such SD files, including structure drawing and rigorous structure validation.

If you would like to discuss this in more detailed, please contact me (Glenn Myatt; glenn.myatt@instem.com).

Reference

  1. In silico insider blog: US FDA collaborations – predicting mutagenicity and beyond…
  2. US FDA poster: Structure-Data File from Sponsors’ Submissions Promotes Rapid Chemical Registration and Computational analyses
  3. US FDA poster: Stop Re-drawing Chemicals! Implementation of Computer-Readable Chemical Structure Format for Drug Impurities
  4. Quick Guide to Creating a Structure-Data File (SD File) for DMF Submissions
  5. Drug Master File (DMF) Submission Resources

Four papers accepted for publication

Over the last couple of weeks, we have received news that four papers, submitted earlier this year to the Journal of Computational Toxicology, have been accepted for publication.

  1. Developing Structure-Activity Relationships for N-Nitrosamine Activity1

This paper outlines N-Nitrosamine carcinogenic potency ranges and describes specific structural features that have clear effects on these ranges. It highlights how N-nitrosamine structure-activity relationships require the investigation of metabolism, DNA binding and DNA repair. Such analysis, based on the presence or absence of structural features, can support read-across for novel N-nitrosamines.

2. In silico approaches in organ toxicity hazard assessment: current status and future needs in predicting liver toxicity2

This paper summarizes the biological mechanisms and processes underlying liver toxicity and describes specific experimental approaches to support the prediction of hepatotoxicity. It discusses the role of in silico approaches and highlights challenges to the adoption of these methods. A proposed framework for the integration of in silico and experimental information is presented.

3. In silico approaches in organ toxicity hazard assessment: current status and future needs for predicting heart, kidney and lung toxicities3

An overview of the key characteristics/mechanisms of heart, lung and kidney toxicity is presented, along with an assessment of how computational methods can be used to understand chemicals that could induce organ toxicity. Obstacles to the development of computational methods predictive of target organ toxicity are discussed.

4. In silico approaches in carcinogenicity hazard assessment: current status and future needs4

The paper summarizes the 10 key characteristics (KCs) of carcinogens and assesses how current in silico methods address each of these KCs. The paper indicates where experimental methods need to be implemented and databases generated. The paper also highlights interactions among the KCs with the different stages of carcinogenesis.

Three of these publications2-4 will appear in a special issue of the journal on the in silico toxicology protocol initiative.

We would like to thank all co-authors for their contribution to these manuscripts.

If you would like to discuss the findings in these manuscripts, please contact me (Glenn Myatt, glenn.myatt@instem.com)

References

1. Arianna Bassan, Vinicius M. Alves, Alexander Amberg, Lennart T. Anger, Scott Auerbach, Lisa Beilke, Andreas Bender, Mark T.D. Cronin, Kevin P. Cross, Jui-Hua Hsieh, Nigel Greene, Raymond Kemper, Marlene T. Kim, Moiz Mumtaz, Tobias Noeske, Manuela Pavan, Julia Pletz, Daniel P. Russo, Yogesh Sabnis, Markus Schaefer, David T. Szabo, Jean-Pierre Valentin, Joerg Wichard, Dominic Williams, David Woolley, Craig Zwickl, Glenn J. Myatt (2021), In silico approaches in organ toxicity hazard assessment: current status and future needs in predicting liver toxicity, Computational Toxicology, 100187. https://doi.org/10.1016/j.comtox.2021.100187

2. Arianna Bassana, Vinicius M. Alves, Alexander Amberg, Lennart T. Anger, Lisa Beilke, Andreas Bender, Autumn Bernal, Mark T.D. Cronin, Jui-Hua Hsieh, Candice Johnson, Raymond Kemper, Moiz Mumtaz, Louise Neilson, Manuela Pavan, Amy Pointon, Julia Pletz, Patricia Ruiz, Daniel P. Russo, Yogesh Sabnis, Reena Sandhu, Markus Schaefer, Lidiya Stavitskaya, David T. Szabo, Jean-Pierre Valentin, David Woolley, Craig Zwickl, Glenn J. Myatt, (2021), In silico approaches in organ toxicity hazard assessment: current status and future needs for predicting heart, kidney and lung toxicities, Computational Toxicology, 100188. https://doi.org/10.1016/j.comtox.2021.100188

3. Kevin P. Cross, David J. Ponting (2021) Developing Structure-Activity Relationships for N-Nitrosamine Activity, Computational Toxicology, 100186. https://doi.org/10.1016/j.comtox.2021.100186

4. Raymond R. Tice, Arianna Bassan, Alexander Amberg, Lennart T. Anger, Marc A. Beal, Phillip Bellion, Romualdo Benigni, Jeffrey Birmingham, Alessandro Brigo, Frank Bringezu, Lidia Ceriani, Ian Crooks, Kevin Cross, Rosalie Elespuru, David M. Faulkner, Marie C. Fortin, Paul Fowler, Markus Frericks, Helga H.J. Gerets, Gloria D. Jahnke, David R. Jones, Naomi L. Kruhlak, Elena Lo Piparo, Juan Lopez-Belmonte, Amarjit Luniwal, Alice Luu, Federica Madia, Serena Manganelli, Balasubramanian Manickam, Jordi Mestres, Amy L. Mihalchik-Burhansa, Louise Neilson, Arun Pandiri, Manuela Pavan, Cynthia V. Rider, John P. Rooney, Alejandra Trejo-Martin, Karen H. Watanabe-Sailor, Angela T. White, David Woolley, Glenn J. Myatt (2021), In Silico Approaches In Carcinogenicity Hazard Assessment: Current Status and Future Needs, Computational Toxicology, in press

US FDA collaborations – predicting mutagenicity and beyond…

This week, we are pleased to welcome Dr. Kevin P. Cross, Instem’s Principal Investigator with U.S. FDA Collaborations and VP of Product Engineering and Production, as a guest contributor to the blog.

Over the last 15 years I have been working closely with the US Food and Drug Administration (FDA) as a principal investigator on several Research Collaboration Agreements (RCAs). We recently extended another 5-year RCA with the Center for Drug Evaluation and Research (CDER) and have additional on-going RCAs with the Center for Food Safety and Applied Nutrition, the Center for Veterinary Medicine, and the Center for Devices and Radiological Health. This has been a long, very productive, and mutually beneficial relationship.

You may have heard of our collaboration to develop new and updated mutagenicity prediction models that are used as part of regulatory approvals for pharmaceutical impurities (ICH M7) and for the assessment of extractables and leachables in medical devices, as well as other applications. In support of this work, we have collaborated on numerous publications documenting these models1 as well as publications outlining the principles and procedures for (Q)SAR assessments that include case studies covering expert review of (Q)SAR results.2,3,4,5

However, you may not be aware of other toxicological endpoints where we have been developing predictive models through the RCAs. This has included the development of new and updated models for many genetic toxicology endpoints,6 organ toxicity, carcinogenicity7, reproductive and developmental toxicity, and neurotoxicity. We also have an on-going collaboration focusing on developing computational translational models that incorporate experimental data in SEND submissions from sponsors alongside traditional QSAR descriptors as part of models to identify clinical liver safety signals using non-clinical data.8 Recently, we have started some new collaborations in the areas of assessing bioactivation (to support the in vitro drug interaction studies guidance9), abuse liability and sex-specific toxicological assessments (for the Office of Women’s Health).

In the past few years, we have also collaborated on the development of an updated version of Leadscope Enterprise that is installed behind the FDA’s firewall at CDER that contains both Leadscope models and databases alongside FDA’s proprietary data. This system supports their computational toxicology group to register new chemicals into their internal chemical dictionary and manage FDA-generated (Q)SAR consultation reports (expert reviewed assessments) that are linked to chemicals in the database. This content is also provided to CDER drug reviewers through a simple and easy-to-use web browser that allows for the searching of the databases and immediate access to these (Q)SAR consultation reports.10

Please feel free to contact me (Kevin Cross; kevin.cross@instem.com) if you would like any additional information.

References

1. Landry, C., Kim, M.T., Kruhlak, N.L., Cross, K.P., Saiakhov, R., Chakravarti, S., Stavitskaya, L., (2019) Transitioning to composite bacterial mutagenicity models in ICH M7 (Q)SAR analyses. Regulatory Toxicology and Pharmacology 109, 104488.

2. Amberg, A., Beilke, L., Bercu, et al., 2016. Principles and procedures for implementation of ICH M7 recommended (Q)SAR analyses. Regulatory Toxicology and Pharmacology 77, 13–24. doi:10.1016/j.yrtph.2016.02.004 

Open access: https://www.sciencedirect.com/science/article/pii/S0273230016300277 

3. Amberg, A., Andaya, R.V., Anger, L.T., et al., 2019. Principles and procedures for handling out-of-domain and indeterminate results as part of ICH M7 recommended (Q)SAR analyses. Regulatory Toxicology and Pharmacology 102, 53–64. doi:10.1016/j.yrtph.2018.12.007 

Open access: https://www.sciencedirect.com/science/article/pii/S0273230018303143 

4. Hasselgren C., et al., (2020) Management of Pharmaceutical ICH M7 (Q)SAR Predictions – The Impact of Model Updates. Regulatory Toxicology and Pharmacology,118, October 2020.

5. Myatt, G.J., Ahlberg, E., Akahori, Y., et al., 2018. In silico toxicology protocols. Regulatory Toxicology and Pharmacology 96, 1–17. doi:10.1016/j.yrtph.2018.04.014 

Open access: https://www.sciencedirect.com/science/article/pii/S0273230018301144 

6. Yoo, J.W., Kruhlak, N.L., Landry, C., Cross, K.P., Sedykh, A., Stavitskaya, L., (2020). Development of improved QSAR models for predicting the outcome of the in vivo micronucleus genetic toxicity assay. Regulatory Toxicology and Pharmacology 113, 104620.

7. Guo, D., Kruhlak, N.L., Stavitskaya, L., Cross, K.P., Bower, D.A., “Characterizing Compound Classes by Rodent Carcinogenicity Tumor Severity and Type”, American College of Toxicology 36th Annual Meeting, November 6-10, 2016, Baltimore, MD.

8. Daniel P. Russo, Kevin P. Cross, Frederic Moulin, Kevin Snyder, Characterizing rat clinical chemistry tests for liver disease phenotypes: a large-scale, cross-study analysis using standardized electronic submission data, in submission.

9. In Vitro Drug Interaction Studies — Cytochrome P450 Enzyme- and Transporter-Mediated Drug Interactions Guidance for Industry https://www.fda.gov/media/134582/download

10. Kevin P. Cross, Marlene T. Kim, Naomi L., Kruhlak, The FDA/CDER Chemical Dictionary, Genetic Toxicity Association Meeting, 2019.

Using in silico approaches to assess abuse liability

In part fueled by the Opioid crisis, the assessment of abuse liability is an area of increased scrutiny. An FDA guidance outlines some recommendations on how to assess abuse potential1.  Several recent publications have illustrated how in silico approaches could be applied to this problem.2,3 Such an assessment could be used as part of early screening to prioritize chemicals, as part of the weight of evidence in assessing a testing strategy and to support regulatory questions.

In collaboration with the FDA, we have also been looking into assessing abuse liability through the development of an abuse liability reference database – an important first step. This database includes chemicals with known abuse liability that are identified as part of national or international registries, such as the United States Drug Enforcement Administration (DEA) scheduled drugs list4 or other publications. Marketed pharmaceuticals that are not on any such list, with no other evidence of abuse potential, are used to classify chemicals as having no evidence of abuse liability.

This database containing positive and negative examples is being used to support both a read-across assessment and to generate structural alerts to profile untested chemicals. By integrating other sources of information, it may be possible to perform a more mechanistic assessment of abuse potential.

Please feel free to contact me (Glenn Myatt; glenn.myatt@instem.com) if you are interested in collaborating on this interesting and emerging topic.

References

  1. FDA Assessment of Abuse Potential of Drugs – Guidance for industry https://www.fda.gov/media/116739/download
  2. Christopher R. Ellis , Rebecca Racz , Naomi L. Kruhlak , Marlene T. Kim , Edward G. Hawkins , David G. Strauss and Lidiya Stavitskaya, Assessing the Structural and Pharmacological Similarity of Newly Identified Drugs of Abuse to Controlled Substances Using Public Health Assessment via Structural Evaluation CLINICAL PHARMACOLOGY & THERAPEUTICS | VOLUME 106 NUMBER 1 | JULY 2019 doi:10.1002/cpt.1418
  3. James W. Firman, Samuel J. Belfield, George Chen, Megan Jackson, Fai Hou Lam, Callum Richmond,  James Smith, Fabian P. Steinmetz, and Mark T. D. Cronin Chemoinformatic Consideration of Novel Psychoactive Substances: Compilation and Preliminary Analysis of a Categorised Dataset. Mol. Inf. 2019, 38, 1800142. DOI: 10.1002/minf.201800142
  4. https://www.dea.gov/drug-information/drug-scheduling

Streamlining ICH M7 analyses with an implemented protocol

A major focus of our work over the last few years has been on the development and publication of in silico toxicology protocols, discussed in previous blog posts1. This has resulted in a paper outlining a framework for such protocols2  as well as two published protocols in the areas of genetic toxicology3 and skin sensitization4. These protocols have been implemented in the Leadscope computational toxicology software to support a fast, defendable, consistent, and fully documented assessment.

We are continually working on the development of new protocols as part of a series of cross-industry working groups. For example, in silico protocols for skin and eye irritation/corrosion as well as endocrine activity are being generated. In addition, many papers outlining the state-of-the-art in the prediction of different toxicological endpoints, such as liver toxicity, are being submitted for publication as part of a special issue on in silico toxicology protocols in the Journal of Computational Toxicology. These newer publications will provide a springboard for the development of new in silico methods and future protocols.

The ICH M7 guideline5 supports the mutagenicity assessment for pharmaceutical impurities and similar approaches are being adopted in many other areas including animal health and pesticide residuals.

Based on the protocol framework, we have now developed an ICH M7 protocol implementation. This pulls together information from the models (statistical QSAR models, expert alerts, cohort-of-concern profilers) and database searches (bacterial mutagenicity, rodent carcinogenicity, acceptable intake limits) which is integrated as part of a documented decision scheme. Figure 1 shows an example of how the new tool summarizes all available information.

Figure 1: Tabular summary of ICH M7 results

In the same manner as the other protocol implementations, it is possible to inspect the underlying decision scheme behind each assessment, perform an expert review (support by a series of guidelines) that may result in modification (such as refuting a (Q)SAR result), as well as documentation of all the results and expert review. Figure 2 shows the decision scheme and expert review guidelines for an individual chemical.

Figure 2: Interactive ICH M7 decision scheme for a single chemical

Please get in touch with me (Glenn Myatt; glenn.myatt@instem.com) if you are interested in discussing this approach to mutagenicity assessment.

References

  1. Can the burden on industry and regulators be reduced? https://insilicoinsider.blog/2020/09/10/can-the-burden-on-industry-and-regulators-be-reduced/
  2. Myatt, G.J., Ahlberg, E., Akahori, Y., et al. (2018) In Silico Toxicology Protocols. Regul. Toxicol. Pharmacol. 98, 1-17. doi:10.1016/j.yrtph.2018.04.014
  3. Hasselgren, C., Ahlberg, E., Akahori, Y., et al. (2019) Genetic toxicology in silico protocol. Regul. Toxicol. Pharmacol.  107, 104403. doi:10.1016/j.yrtph.2019.104403 
  4. Johnson, C., Ahlberg, E., Anger, L.T., et al. (2020) Skin sensitization in silico protocol. Regul. Toxicol. Pharmacol.  116, October 2020, 104688. doi: 10.1016/j.yrtph.2020.104688 
  5.  ICH M7, 2017 (R1) (2017) Assessment and control of DNA reactive (mutagenic) impurities in pharmaceuticals to limit potential carcinogenic risk. https://database.ich.org/sites/default/files/M7_R1_Guideline.pdf

An expert review of potentially reactive features

In several recent posts1,2 we highlighted the usefulness of an expert review of potentially reactive features. This is particularly important when an out-of-domain result is returned or when an area of the test chemical is not being considered by the computational model. The US Food and Drug Administration (FDA) recently published a paper showing that expert knowledge plays an important role in increasing the reliability of (Q)SAR predictions3. Software to support various expert review methodologies is needed to facilitate thorough computational assessments.  

Examining how often different substructural features present in the test chemical appear in historical collections of chemicals with experimental results will tell us whether such a feature is potentially problematic (i.e., there are more toxic examples than you would generally expect) or whether there is no concern (i.e., the proportion of toxic chemicals is either similar to or less than the proportion observed in the whole database). The blog “The use of chemical analogs in expert reviews”2 includes some nice examples.

The Leadscope products make use of a dictionary of over 27,000 substructural fragments often described as the ‘Leadscope feature hierarchy’4. Matching any of these predefined structural features, that are also present in your test chemical, against a database is a great starting point. However, there may still be questions about specific chemical fragments that are not included in this list.

Over the last year, we have been working hard to implement a brand-new chemical structure drawing package which is being integrated into the Leadscope tools. Although one of the more common applications of this tool will be to draw a chemical structure on which to apply computational models, the tool is also ideal for the assessment of ad hoc potentially reactive features.

Please get in touch with me (Glenn Myatt, glenn.myatt@instem.com) if you would like to discuss this approach in more detail.

References

  1. How an expert-review could be used to resolve out-of-domains
  2. The use of chemical analogs in expert reviews
  3. Jayasekara, S et al., Assessing the impact of expert knowledge on ICH M7 (Q)SAR predictions. Is expert review still needed?, Regulatory Toxicology and Pharmacology, 125,  2021, 105006
  4. Roberts G, et al., Leadscope: Software for Exploring Large Sets of Screening Data. J. Chem. Inf. Comput. Sci., 2000, 40, 1302-1314.