by George Fitzgerald, Ph.D.
Today’s cosmetics involve a wide array of functional additives, as well as ones that simply improve consumer appeal. These may impact, for example, the absorption of actives by skin, improve rinsability or extend wear time. How does a company cost effectively optimize these factors to distinguish itself from competitors, expand market share and drive top-line growth?
Cosmetics manufacturers would do well to take a page from the playbook of the pharmaceutical and petroleum industries. Both have embraced emerging technologies such as molecular modeling and predictive analytics to streamline and accelerate experimental progress. This article discusses how innovative scientific informatics solutions already proven in other research-driven enterprises can be used to slash costs, optimize formulations, improve quality control and ultimately speed the development of new products.
Data Integration, Simplified
Cosmeceuticals development requires the investigation of a much broader range of compounds than traditional cosmetics. In order to identify effective and marketable actives, researchers need to understand factors as diverse as transdermal penetration, color fastness, shelf life and skin sensitivity. This involves a massive volume of information—everything from thousands of possible formulation ingredients to chemical models, biological assays, previous experimental results and more. But all this information is not going to be particularly useful until it can be collected, tracked and analyzed as a cohesive whole.
Data integration may seem like a simple concept, but in a scientific environment it’s actually quite complex because of the wide diversity of data formats, the sheer number of instruments that need to be accommodated, and the many different locations within the R&D organization where research intelligence may be hidden. Researchers can easily spend countless hours finding needed information, preparing data for analysis, and collating, formatting and distributing results—time that could be more profitably spent on product innovation.
Fortunately, next-generation service-oriented technologies are changing this by enabling a more unified approach to managing complex scientific information. For example, a Web services-based foundation for scientific informatics can support the integration of multiple sources of information in a “plug and play” environment. This allows project stakeholders to: create automated workflows that streamline experimental progress; conduct advanced modeling, analysis and reporting across different data sets; institutionalize best practices; and more easily collaborate without the time and expense involved in writing custom software.
A smarter, more integrated approach to tracking organizational knowledge additionally enables companies to build on the valuable, yet time consuming research they’ve already invested in, rather than start from scratch with every project. When individual contributors can more easily archive, search and compare information from past experiments with current projects, huge efficiency gains can be made.
Predictive Modeling—A Faster Path to Innovation
Predictive science is an especially important tool when little experimental data exists. This involves the use of sophisticated theoretical methods to compute results even when experimental data sets are unavailable. Used widely in pharmaceutical research, software-enabled scientific modeling and analytic techniques make it possible for researchers to design chemical compounds or screen for properties like effectiveness and toxicity in silico. With these kinds of tools, researchers can quickly narrow the search for promising leads and optimize formulation recipes before running costly and time consuming trial and error-type experiments in the lab.
As the name suggests, molecular modeling predicts properties of materials or blends at the molecular level. In cosmeceutical research, this is especially useful in understanding dermal delivery (i.e., how active ingredients will penetrate the skin)—a key point of product differentiation.
Molecular modeling has long been able to predict the efficacy of small molecule active ingredients, and recent advances have made it possible to create realistic models of the lipid layers found in human skin. With these models, researchers can probe the dermal delivery process for a wide range of skin types, active ingredients and external conditions in order to identify compounds that deliver optimal absorption. How, for example, does pH affect skin absorption? Given two comparable ingredients, which absorbs more quickly? With the ability to investigate a broad range of compounds computationally, researchers can explore far more options than they could through lab experiments alone.
While molecular models are primarily used to identify and design lead compounds from scratch, numerical (or phenomenological) models use existing experimental data as the basis to quickly and reliably predict the behavior of new materials or mixtures. These are useful for both formulation optimization and in product safety, specifically in irritancy or toxicity screening.
Formulations are key to the success of virtually every product in cosmetics. Active ingredients must be mixed with pigments and inactive substrates, for example, so they include textures and fragrances that appeal to consumers. Predictive analytics provides a way of designing formulations on the basis of the available data so costs can be kept in check. For example, models can be used to screen thousands of virtual formulations, with the goal of finding the optimum combination of ingredients. Only the most promising formulations are then subjected to experimental screening, reducing the number of laboratory experiments required.
And, as any researcher knows, there are countless variables that can impact a formulation’s safety and trigger undesirable side effects. With modeling tools, organizations can also eliminate chemical combinations that could potentially cause skin sensitivity or other issues. Doing this virtually not only saves a great deal of time and money, it also helps companies more easily comply with regulatory standards like REACH, the European Union mandate that sets safety standards on chemical substances.
Both molecular and empirical modeling make it possible for R&D teams to design and screen far more molecules, mixtures or recipes in silico than they could through experimentation alone. Researchers can perform calculations on all of their leads and subject only the most promising—say the top 25 percent—to experimentation. This has the potential to reduce experimental costs by about 75 percent, which is a significant savings. And the time saved in the lab helps companies get competitive new products to market faster.
Putting lipstick or skin cream on a beautiful model is not the only way to make a cosmeceutical product look good. Sophisticated computational models and other emerging solutions for scientific informatics are transforming research innovation and enabling cosmetics companies to develop new products that have high consumer appeal. By enabling organizations to engage in predictive science at all stages of the research and development process, these technologies offer great potential to lower overall costs, reduce time-to-market, increase the productivity of personnel and positively impact the bottom line.
George Fitzgerald, Ph.D., is a recognized modeling and simulation specialist with experience in both drug discovery as well as research areas involving chemicals and materials. Today at Accelrys (Accelrys.com), he is involved in developing molecular modeling and informatics solutions for the chemicals, materials, CPG and related industries. He spent several years at Biosym Technologies developing methods for computational drug discovery. In the 1990s, he worked at Cray Research, where he worked closely with customers to develop high-performance solutions for the chemicals industry. He is the author of more than 50 peer-reviewed publications and has been invited to present at numerous scientific and industrial conferences. Fitzgerald received his B.S in chemistry from Case Western Reserve University and his Ph.D. from the University of California, Berkeley, where he studied molecular modeling. His blog can be found at http://blog.accelrys.com/author/george/