Michael Doyle Ph.D., Accelrys09.30.10
From anti-aging skin crèmes to extended wear lipstick, today’s top selling cosmetics all have one thing in common: performance.
As more and more personal care companies seek to grow their market share by developing products with active, clinically-claimed ingredients, the line between cosmetics and pharmaceuticals is becoming increasingly blurred. Today’s topical pharmaceutical material is tomorrow’s volume anti-aging cream. The goal is to identify differentiating additives that produce demonstrable effects that will entice consumers to buy. As a result, R&D organizations must be able to investigate a much broader range of compounds and formulation systems than ever before. And in an industry where getting to market first is paramount, they must also deliver innovations as efficiently and cost effectively as possible.
Developing safe, stable, active and market-relevant cosmeceutical systems requires extremely sophisticated analysis capabilities. Researchers need to understand factors as diverse as delivery, image, material shelf life and skin sensitivity. They also need to be sure that a system can be mass-produced in multiple locations both rapidly and affordably. It is prohibitively expensive and time consuming to explore all the potential variables that may impact a system through empirical or trial-and-error lab experimentation alone. Furthermore, experimental results are not always optimal.
To address the challenges presented by cosmeceutical product development, personal care companies are applying data mining and knowledge analysis processes—such as scientific informatics, virtual chemistry and formulation analysis techniques—to augment their laboratory efforts. These approaches were proven in the pharmaceutical industry, where they enabled companies to streamline and accelerate product design, testing and optimization.
Here are some ways that R&D organizations focusing on cosmetics can reduce the number of experiments needed, slash costs and speed innovation.
In-Silico Modelling
Leaf through any glossy magazine and look at the cosmetics advertisements. It’s clear that to be competitive, products must have some level of provable effectiveness for the consumer, such as anti-aging, color stability or UVA/UVB protection. As a result, cosmetic researchers are focusing more on designing compounds that do more than just sit on the surface of the skin. Active ingredients that can penetrate layers of the epidermis or bind to a particular protein are now essential to competitive differentiation. But pinpointing the right compound that will produce a desired cosmeceutical effect can be like finding a needle in a haystack, as there are thousands of possible molecular combinations that might potentially unlock the best lead. Traditional, trial and error-type approaches to discovery often require many expensive one-off experiments, which are both cost and time prohibitive. These challenges are exacerbated by the increasing need to “customize” cosmetic systems to meet the unique and continually changing needs of specific geographic regions, climates and cultures, or to address other factors such as skin type and personal health.
Technology that facilitates complex molecular simulation presents a compelling alternative to lab experimentation. Used widely in pharmaceutical research, software-enabled scientific modelling and analytic techniques make it possible for researchers to design molecular compounds in silico, applying a more streamlined, predictive approach to discovery. Instead of using trial and error, researchers can take advantage of molecular models to more easily explore a broad range of compound candidates, weed out those that won’t hit the mark, and quickly narrow the search for the most promising leads. This computational approach delivers not just leads but important, detailed information about how those molecular-level features that distinguish successful products actually work, so that marketing claims can be supported.
For example, modelling can be used to better understand dermal delivery—one of the primary areas of product differentiation in cosmeceuticals. By leveraging computational tools to build realistic molecular models of the lipid layers found in human skin, researchers can probe the dermal delivery process across a wide range of skin types, active ingredients and external conditions in order to identify compounds that deliver optimal efficacy. How, for example, does pH affect skin absorption? Given two comparable ingredients, which absorbs more quickly? Which molecular combinations are most effective in blocking UVA and UVB sun rays?
Here’s a recent real-world illustration: a leading global cosmetics brand launched an anti-aging serum last year that marked the company’s first product developed with the help of genomic information and analysis approaches.
With the ability to investigate a broad range of compounds computationally, researchers can explore far more options than would be possible through lab experiments alone. They can figure out what doesn’t work before investing too heavily in synthesis, screening and testing. And, if an organization can narrow its synthesis candidates down from thousands to hundreds or even less, the discovery process will be correspondingly faster and far more cost effective.
Virtual Screening
There are countless variables that can compromise a cosmetic product’s stability or trigger undesirable side effects. For example, because formulations include a number of both active and inactive ingredients, researchers must determine how each potential combination will behave under different conditions. Will the ingredients separate after sitting on the shelf for a period of time or break down when the temperature gets too high or too low? Even more importantly, is the formulation safe to use? Testing for things such as skin sensitivity is critical, especially as the introduction of active ingredients into cosmetics opens the door to more stringent regulation.
But, just as synthesizing every possible active compound is too costly and time consuming during the discovery phase of R&D, so too is testing every ingredient (and all the potential combinations of ingredients) that might impact morphology or safety. Where molecular models are primarily used to identify and design lead compounds in the absence of experimental results, analytic models can leverage existing data sets to quickly and reliably predict the behavior of formulations and mixtures. These models can screen thousands of formulations virtually, so that researchers can pinpoint optimal recipes that are safe and stable, as well as identify potential problem areas, before expensive and painstaking lab testing.
Furthermore, as product testing methods come under increasing scrutiny due to regulatory standards such as REACH, there is a pressing need for screening approaches that help companies comply with new regulations while simultaneously reducing the need for animal testing. Predictive models are especially suited to dealing with this challenge.
For example, pharmaceutical companies have routinely leveraged quantitative structure-activity prediction (QSAR) models to perform rapid assessments of chemical toxicity based solely on a chemical’s molecular structure. QSAR models can be used for a wide range of screens including physical/chemical, environmental fate, ecotoxicity, toxicity, mutagenicity and others. These same techniques can easily be applied to cosmeceutical testing, and can also be automated so that thousands of compounds can be piped through a model that is updated continuously with the latest experimental results, in order to yield the best possible predictions.
The bottom line is that computational models offer a virtual alternative to stability and safety screening, reducing the need for multiple lab experiments. This helps researchers keep costs in check, and offers a huge advantage in terms of speeding development cycle times.
Shared Knowledge
The knowledge that drives product innovation can come from many sources. In addition to the data generated from current experiments (both through the in silico approaches described above as well as in the laboratory), researchers must incorporate information from previous projects, the scientific literature and in-house and publicly available databases. Most product development does not start from scratch. Researchers often begin with a product “chassis”—perhaps a basic skin cream formulation—and then add active attributes such as anti-aging or skin lightening capabilities that might appeal to the target market. Innovation usually results from a combination of experimentation and building on what came before.
A Huge Knowledge Base
The good news is that the available knowledge base that researchers can tap into when developing compelling new products is huge—and that’s also the bad news. A typical personal care product development project may span many scientific disciplines and include information related to thousands of possible compounds and formation ingredients, high throughput experimental results, historical project information and more. This information is also often spread across a diverse array of formats and proprietary systems, such as text documents saved in an electronic lab notebook or images generated by a microscope. Disciplinary silos create further complications, locking information within a particular department or research group (biology experiments may be conducted independently of chemistry experiments or toxicology testing, for example). As a result, stakeholders may spend countless hours tracking down what they need or they may simply miss critical knowledge entirely, burning resources on redundant experiments.
Extracting maximum value from massive quantities of scientific data requires both the ability to integrate disparate information sources as well as quickly find the content most relevant to the research problem at hand. Being able to analyze data as a cohesive whole, especially across different areas of specialization, allows researchers to make important connections that otherwise would have been missed. They also must be able to access specific information—a skin cell image or an existing formulation recipe—without a lot of hassle.
Fortunately, new advances in service-oriented architecture and collaborative technologies are enabling a more unified approach to managing complex scientific data. For example, a web services-based foundation for scientific informatics can bring together multiple sources of data in a “plug and play” environment. This allows project stakeholders to create automated workflows that streamline experimental progress and conduct modelling, analysis and reporting across different data sets. Additionally, when information is all in one place, it can more easily be searched. Through this “shared knowledge base,” organizations can enable richer collaboration, faster innovation and more effectively capitalize on all their information resources.
Consider the example of a company working to introduce a sunscreen formula originally developed for European consumers to Asian markets. While chemists can likely build on the existing formulation, they also need to ensure that that the Asian version is optimized to meet the unique tastes and requirements of the target consumer. As a result, there are a number of variables that must be considered.
Variables to Consider
First, there’s the original formulation:How did it test under different temperature and environmental conditions? What are the solubility profiles? Can this mixture be modified easily in order to change its color or texture in order to address the preferences and realities of the new market?
There are also local considerations: How does the climate and geography impact the efficacy of the product’s UV blocker? What about its shelf life? Will Asian skin (due to differences in genetics or diet) absorb the product differently? Can additional ingredients (such as an active that will enhance the product’s ability to penetrate the skin) be mixed safely and effectively with the original formulation?
With a system in place that enables development teams to easily find, access and use any type of existing R&D knowledge related to the sunscreen formulation (as well as information on suppliers, processing data, test results and more), the cosmetics organization can build on existing innovation to reach new markets.
For instance, data about the original formulation’s solubility profile and computed irritancy indexes can be used to create models that predict how the product might perform on different skin types or with different active ingredients added. Additives that may cause an undesirable reaction or that are simply ineffective can be ruled out before clinical trials are required. This can mean the difference between having to test four or five ingredients in vivo instead of 20—which can reduce costs significantly and potentially slash weeks from the development cycle.
A smarter, more integrated approach to tracking organizational knowledge enables companies to build on the valuable, yet time consuming, research they’ve already invested in, rather than start from scratch with every project. Additionally, they can maximize their investments in technologies that generate large volumes of data, such as high throughput experimentation (HTE) equipment.
Fiercely competitive and fast paced, the personal care industry relies on product innovation to maintain existing market share and acquire new customers, drive profitability and stimulate business growth. Increasingly, this means developing cosmetic products that include active ingredients, adding a new layer of complexity to the R&D process. As the discovery, screening and data management requirements of cosmetics companies more closely resembles those of their pharmaceutical counterparts, organizations can take advantage already proven technologies—including predictive modelling, advanced analytics and scientifically-focused informatics solutions—to reduce costs and speed time-to-market.
As more and more personal care companies seek to grow their market share by developing products with active, clinically-claimed ingredients, the line between cosmetics and pharmaceuticals is becoming increasingly blurred. Today’s topical pharmaceutical material is tomorrow’s volume anti-aging cream. The goal is to identify differentiating additives that produce demonstrable effects that will entice consumers to buy. As a result, R&D organizations must be able to investigate a much broader range of compounds and formulation systems than ever before. And in an industry where getting to market first is paramount, they must also deliver innovations as efficiently and cost effectively as possible.
Developing safe, stable, active and market-relevant cosmeceutical systems requires extremely sophisticated analysis capabilities. Researchers need to understand factors as diverse as delivery, image, material shelf life and skin sensitivity. They also need to be sure that a system can be mass-produced in multiple locations both rapidly and affordably. It is prohibitively expensive and time consuming to explore all the potential variables that may impact a system through empirical or trial-and-error lab experimentation alone. Furthermore, experimental results are not always optimal.
New lipstick launches are faster with predictive modelling. |
Here are some ways that R&D organizations focusing on cosmetics can reduce the number of experiments needed, slash costs and speed innovation.
In-Silico Modelling
Leaf through any glossy magazine and look at the cosmetics advertisements. It’s clear that to be competitive, products must have some level of provable effectiveness for the consumer, such as anti-aging, color stability or UVA/UVB protection. As a result, cosmetic researchers are focusing more on designing compounds that do more than just sit on the surface of the skin. Active ingredients that can penetrate layers of the epidermis or bind to a particular protein are now essential to competitive differentiation. But pinpointing the right compound that will produce a desired cosmeceutical effect can be like finding a needle in a haystack, as there are thousands of possible molecular combinations that might potentially unlock the best lead. Traditional, trial and error-type approaches to discovery often require many expensive one-off experiments, which are both cost and time prohibitive. These challenges are exacerbated by the increasing need to “customize” cosmetic systems to meet the unique and continually changing needs of specific geographic regions, climates and cultures, or to address other factors such as skin type and personal health.
Technology that facilitates complex molecular simulation presents a compelling alternative to lab experimentation. Used widely in pharmaceutical research, software-enabled scientific modelling and analytic techniques make it possible for researchers to design molecular compounds in silico, applying a more streamlined, predictive approach to discovery. Instead of using trial and error, researchers can take advantage of molecular models to more easily explore a broad range of compound candidates, weed out those that won’t hit the mark, and quickly narrow the search for the most promising leads. This computational approach delivers not just leads but important, detailed information about how those molecular-level features that distinguish successful products actually work, so that marketing claims can be supported.
The cosmetics industry could learn a lesson from Big Pharma. |
For example, modelling can be used to better understand dermal delivery—one of the primary areas of product differentiation in cosmeceuticals. By leveraging computational tools to build realistic molecular models of the lipid layers found in human skin, researchers can probe the dermal delivery process across a wide range of skin types, active ingredients and external conditions in order to identify compounds that deliver optimal efficacy. How, for example, does pH affect skin absorption? Given two comparable ingredients, which absorbs more quickly? Which molecular combinations are most effective in blocking UVA and UVB sun rays?
Here’s a recent real-world illustration: a leading global cosmetics brand launched an anti-aging serum last year that marked the company’s first product developed with the help of genomic information and analysis approaches.
With the ability to investigate a broad range of compounds computationally, researchers can explore far more options than would be possible through lab experiments alone. They can figure out what doesn’t work before investing too heavily in synthesis, screening and testing. And, if an organization can narrow its synthesis candidates down from thousands to hundreds or even less, the discovery process will be correspondingly faster and far more cost effective.
Virtual Screening
There are countless variables that can compromise a cosmetic product’s stability or trigger undesirable side effects. For example, because formulations include a number of both active and inactive ingredients, researchers must determine how each potential combination will behave under different conditions. Will the ingredients separate after sitting on the shelf for a period of time or break down when the temperature gets too high or too low? Even more importantly, is the formulation safe to use? Testing for things such as skin sensitivity is critical, especially as the introduction of active ingredients into cosmetics opens the door to more stringent regulation.
But, just as synthesizing every possible active compound is too costly and time consuming during the discovery phase of R&D, so too is testing every ingredient (and all the potential combinations of ingredients) that might impact morphology or safety. Where molecular models are primarily used to identify and design lead compounds in the absence of experimental results, analytic models can leverage existing data sets to quickly and reliably predict the behavior of formulations and mixtures. These models can screen thousands of formulations virtually, so that researchers can pinpoint optimal recipes that are safe and stable, as well as identify potential problem areas, before expensive and painstaking lab testing.
Furthermore, as product testing methods come under increasing scrutiny due to regulatory standards such as REACH, there is a pressing need for screening approaches that help companies comply with new regulations while simultaneously reducing the need for animal testing. Predictive models are especially suited to dealing with this challenge.
For example, pharmaceutical companies have routinely leveraged quantitative structure-activity prediction (QSAR) models to perform rapid assessments of chemical toxicity based solely on a chemical’s molecular structure. QSAR models can be used for a wide range of screens including physical/chemical, environmental fate, ecotoxicity, toxicity, mutagenicity and others. These same techniques can easily be applied to cosmeceutical testing, and can also be automated so that thousands of compounds can be piped through a model that is updated continuously with the latest experimental results, in order to yield the best possible predictions.
The bottom line is that computational models offer a virtual alternative to stability and safety screening, reducing the need for multiple lab experiments. This helps researchers keep costs in check, and offers a huge advantage in terms of speeding development cycle times.
Shared Knowledge
The knowledge that drives product innovation can come from many sources. In addition to the data generated from current experiments (both through the in silico approaches described above as well as in the laboratory), researchers must incorporate information from previous projects, the scientific literature and in-house and publicly available databases. Most product development does not start from scratch. Researchers often begin with a product “chassis”—perhaps a basic skin cream formulation—and then add active attributes such as anti-aging or skin lightening capabilities that might appeal to the target market. Innovation usually results from a combination of experimentation and building on what came before.
A Huge Knowledge Base
The good news is that the available knowledge base that researchers can tap into when developing compelling new products is huge—and that’s also the bad news. A typical personal care product development project may span many scientific disciplines and include information related to thousands of possible compounds and formation ingredients, high throughput experimental results, historical project information and more. This information is also often spread across a diverse array of formats and proprietary systems, such as text documents saved in an electronic lab notebook or images generated by a microscope. Disciplinary silos create further complications, locking information within a particular department or research group (biology experiments may be conducted independently of chemistry experiments or toxicology testing, for example). As a result, stakeholders may spend countless hours tracking down what they need or they may simply miss critical knowledge entirely, burning resources on redundant experiments.
Extracting maximum value from massive quantities of scientific data requires both the ability to integrate disparate information sources as well as quickly find the content most relevant to the research problem at hand. Being able to analyze data as a cohesive whole, especially across different areas of specialization, allows researchers to make important connections that otherwise would have been missed. They also must be able to access specific information—a skin cell image or an existing formulation recipe—without a lot of hassle.
Fortunately, new advances in service-oriented architecture and collaborative technologies are enabling a more unified approach to managing complex scientific data. For example, a web services-based foundation for scientific informatics can bring together multiple sources of data in a “plug and play” environment. This allows project stakeholders to create automated workflows that streamline experimental progress and conduct modelling, analysis and reporting across different data sets. Additionally, when information is all in one place, it can more easily be searched. Through this “shared knowledge base,” organizations can enable richer collaboration, faster innovation and more effectively capitalize on all their information resources.
Consider the example of a company working to introduce a sunscreen formula originally developed for European consumers to Asian markets. While chemists can likely build on the existing formulation, they also need to ensure that that the Asian version is optimized to meet the unique tastes and requirements of the target consumer. As a result, there are a number of variables that must be considered.
Variables to Consider
First, there’s the original formulation:How did it test under different temperature and environmental conditions? What are the solubility profiles? Can this mixture be modified easily in order to change its color or texture in order to address the preferences and realities of the new market?
There are also local considerations: How does the climate and geography impact the efficacy of the product’s UV blocker? What about its shelf life? Will Asian skin (due to differences in genetics or diet) absorb the product differently? Can additional ingredients (such as an active that will enhance the product’s ability to penetrate the skin) be mixed safely and effectively with the original formulation?
With a system in place that enables development teams to easily find, access and use any type of existing R&D knowledge related to the sunscreen formulation (as well as information on suppliers, processing data, test results and more), the cosmetics organization can build on existing innovation to reach new markets.
For instance, data about the original formulation’s solubility profile and computed irritancy indexes can be used to create models that predict how the product might perform on different skin types or with different active ingredients added. Additives that may cause an undesirable reaction or that are simply ineffective can be ruled out before clinical trials are required. This can mean the difference between having to test four or five ingredients in vivo instead of 20—which can reduce costs significantly and potentially slash weeks from the development cycle.
A smarter, more integrated approach to tracking organizational knowledge enables companies to build on the valuable, yet time consuming, research they’ve already invested in, rather than start from scratch with every project. Additionally, they can maximize their investments in technologies that generate large volumes of data, such as high throughput experimentation (HTE) equipment.
Fiercely competitive and fast paced, the personal care industry relies on product innovation to maintain existing market share and acquire new customers, drive profitability and stimulate business growth. Increasingly, this means developing cosmetic products that include active ingredients, adding a new layer of complexity to the R&D process. As the discovery, screening and data management requirements of cosmetics companies more closely resembles those of their pharmaceutical counterparts, organizations can take advantage already proven technologies—including predictive modelling, advanced analytics and scientifically-focused informatics solutions—to reduce costs and speed time-to-market.