As cosmetic and fragrance ingredient portfolios continue to expand, safety assessors face growing pressure to evaluate increasingly diverse chemistries without relying on animal testing as legislation phasing out the practice gains momentum.
At the same time, regulators and stakeholders like non-profit ChemFORWARD continue to call for greater transparency and consistency in ingredient safety, while traditional methods for grouping and classifying ingredients based largely on structural similarity can struggle to account for subtle differences.
To address these challenges, the Research Institute for Fragrance Materials (RIFM) has developed a signature-based chemical grouping framework that links structural features with mechanistic toxicology considerations.
In this CosmeticsDesign Q&A we spoke with Jake Muldoon, PhD, Senior Associate Scientist, Chemistry, and Holger Moustakas, PhD, Senior Scientist and lead of RIFM’s Chemistry Program, about how the framework works in practice, its implications for regulatory confidence, and its potential application across cosmetics and fragrance safety assessment.
CDU: What problem does this signature-based grouping approach solve for non-animal safety assessments, and why is it important for the cosmetics and fragrance industry?
Jake Muldoon (JM): One of the biggest challenges in non-animal safety assessment is ensuring chemicals are grouped for scientifically meaningful reasons. Traditional grouping methods primarily rely on broad structural similarity, which can overlook subtle but toxicologically important differences in how ingredients interact with biological systems.
This is especially true when dealing with the extensive, diverse libraries typical of fragrance and cosmetic ingredients.
Our signature-based approach adds mechanistic context. Instead of simply grouping chemicals that look similar, we consider how structural features influence biological activity: reactivity, metabolism, or other mechanisms underlying potential adverse effects.
The result is a more reliable and reproducible foundation for non-animal safety evaluations. It elevates the scientific fidelity of read-across, making decisions more representative of how chemicals actually behave.
Holger Moustakas (HM): This is crucial in cosmetics and fragrance, where materials require rapid, consistent evaluation and where new chemistries are continually introduced. We need a method that is both efficient and scientifically defensible. SAGs deliver that clarity.
They ensure chemicals are grouped for the right reasons, grounded in mechanistic toxicology rather than subjective expert judgment. Ultimately, this supports more robust, more consistent, and fully non-animal safety decisions, helping the sector advance toward higher-confidence, next-generation assessment practices.
CDU: How do Indicator Phrases (IPs) and Structure-Activity Groups (SAGs) work together to make chemical grouping more transparent and scientifically robust compared to traditional methods?
HM: SAGs and Indicator Phrases were designed to solve different but complementary problems. SAGs cluster chemicals based on structural features known (or strongly predicted) to influence biological activity.
Each SAG represents a mechanistic “family” in which related structural traits signal similar safety concerns or metabolic behavior.
Indicator Phrases translate those mechanistic expectations into concise descriptions. They capture the structural features that influence reactivity patterns, metabolic pathways, and hazards associated with specific functional groups. So, while SAGs define who belongs in a group, IPs explain why the group matters toxicologically.
JM: Together, these tools replace intuition-based grouping with a transparent, criteria-driven rationale. Anyone reviewing an assessment—whether a toxicologist, formulator, safety assessor, or regulator—can see not only which chemicals were grouped together, but the mechanistic reasoning behind it.
This clarity improves reproducibility and enhances regulatory confidence. With explicit structural and mechanistic justification, read-across becomes more predictable, more consistent, and easier to evaluate objectively. In short, SAGs and IPs bring structure and defensibility to a historically opaque process.
CDU: Your system has assessed over 6,000 fragrance chemicals. How scalable is this approach for other sectors like cosmetics and what adaptations would be needed?
JM: Scalability was a core design principle. While we applied the framework first to fragrance ingredients, the underlying concepts (linking structural features to mechanistic activity) apply across chemical sectors.
Many cosmetic ingredients share similar functional groups or structural motifs, allowing them to integrate into this system with minimal adjustment. The broader and more diverse the ingredient palette becomes, the more essential a mechanistically informed grouping system like SAGs becomes for efficient prioritization and safety evaluation.
HM: Importantly, scalability isn’t just about dataset size. It’s about adaptability. As new chemistries emerge, experts can expand or refine SAGs and IPs, making the framework well-suited for integration with computational tools, decision-support systems, and even machine-learning models that rely on consistent, well-annotated data structures.
CDU: What practical benefits can manufacturers expect from using SAGs for read-across and prioritization in safety testing? Does it help streamline regulatory compliance and reduce time-to-market?
HM: Yes. SAGs help manufacturers prioritize testing strategies more effectively by organizing ingredients into mechanistically meaningful groups. This ensures that scientific and financial resources, whether for in vitro testing, computational modeling, or literature review, are focused on areas where new data will have the greatest impact.
When navigating regulatory environments that increasingly expect strong justification for non-animal approaches, having a validated, transparent rationale becomes a major advantage.
JM: Because SAGs and IPs explicitly communicate the scientific basis for grouping, read-across becomes more predictable and more defensible. This reduces the iteration cycle often required with regulators, speeding up assessments and lowering the risk of delays.
For manufacturers, the result is a clearer regulatory pathway, fewer questions during review, and ultimately a faster time-to-market, while maintaining rigorous safety standards. It’s a more efficient, forward-looking approach that supports innovation without compromising consumer protection.
