Are my (bio)pharmaceutical assay performances reliable? Only probability of success counts!
The Global Statistics and Data Science (GSDS) team at PharmaLex delivers analytical support for the development of medical products–from early research through pre-clinical to late phase studies, plus manufacturing and health economics.
We support our clients every step of the way–from discovery and non-clinical requirements to early strategic planning, clinical development, CMC development, regulatory submissions and post-approval/post-launch maintenance.
Within scientific and statistical contexts, we offer interpretation, answers and data science solutions that help you make the best decisions for your business.
Partnering with clients – maintaining relationships
By partnering with our clients and working as a team, we make each other stronger. We have nurtured long term relationships with many of our clients and look forward to forging strong partnerships with many new ones.
By embracing the power of teamwork, we alleviate your short-term work overload, provide niche expertise and become your trusted go-to for all complex data analytics needs.
Combining talent – Adding value
Without a doubt, our employees are an invaluable asset. This group of (Bayesian) statisticians, data scientists, computational biologists, statistical geneticists, epidemiologists, bioengineers, pharmacists, and programmers offers an outstanding level of skill and expertise.
Stretching across many diseases, our combined experience supports both target identification and the development of drugs, biologics and medical devices at every phase.
PharmaLex is recognized as a global leader in the application of Bayesian methodology, biomarker statistics and CMC statistics. Many of our scientists have wet-lab experience which helps them tackle challenges in biological data. We speak and operate in many languages, including SAS, R, STAN, JAGS and Python. Our deep regulatory experience also helps you navigate all statistical facets of study design and regulatory strategy.
Analytics
The Power of Smart Analytics
With recent advances in science and research methodologies, the sheer volume of information at hand has increased the importance of statistics and data science.
Applied correctly, statistics and data science can help:
- Identify or validate new drug targets by means of statistical genomics, bioinformatics and computational biology techniques
- Inform better decisions in clinical development (e.g. more efficient trial design, improved patient profiles)
- Apply prior knowledge of Bayesian statistics to reduce future costs and save time
- Realize the promise of precision medicine through biomarkers and diagnostics
- Lead to safer, more-robust manufacturing by using the quality-by-design concept
- Unlock data insights by using advanced methods like artificial intelligence and machine learning
- Progress the digitalization of the pharmaceutical industry
OUR SERVICES
Biostatistics is the application of statistical thinking and techniques to problems in biology and health. It has long been a crucial component of drug and medical device development, since it enables the scientific distinction between events that are causally associated with an intervention, and those that happen by mere chance. In biopharmaceutical work, tools like survival analysis, longitudinal data analysis, and generalized linear and nonlinear modeling have long been considered core tools. Such investigations can be for the purposes of discovery (as when we are looking for potential associations that merit further investigation), or confirmation (as when we evaluate the results of a clinical trial in a regulated setting).
Traditional biostatistical methods rely on what are called p-values to indicate the level of surprise in a statistical result; a small p-value indicates a result that would be unlikely to have occurred if in fact no underlying scientific relationship existed. While these methods continue to be useful, they are often inadequate in complex, high-dimensional analytic settings. By contrast, Bayesian statistical methods offer a way to obtain direct probability statements about any statistical quantity of interest (say, the probability a drug is safe and effective). While they require more sophisticated programming and computational methods, their increased flexibility is often a boon in many settings, both in discovery and confirmation. They can also be used sequentially as more data become available, in a “learn and confirm” paradigm useful for multiphase clinical trials.
PhamaLex has assembled a world-class team of Bayesian statisticians skilled in a wide variety of areas, including
- Discovery, preclinical and developmental statistics
- Chemistry, Manufacturing and Control (CMC)
- Pharmacokinetic/Pharmacodynamic analysis
- Adaptive Clinical Trial Design (basket trials, umbrella trials, etc)
- Adaptive Incorporation of Auxiliary Information (historical data, registry data, published information, real world data, expert opinion)
- Biomarker statistics
- Meta-analysis and Network Meta-Analysis
- Post-market Studies
Over the past five years, the use of Bayesian and other nontraditional approaches to complex innovative design (CID) has become quite common in the design and analysis of clinical trials. In the United States, this growth has been fostered by the 21st Century Cures Act and PDUFA VI, which encourage the use of novel statistical methods for combining disparate data sources. This increased interest has been especially acute in medical devices and in rare and pediatric drug approvals, but is now also spreading to all areas of regulatory science. Incorporating information from both randomized (e.g., historical trial data) and nonrandomized (e.g., disease registry and other real world evidence) sources offers the potential to dramatically reduce the cost, timeliness, and ethical hazard of clinical trials.
PharmaLex has extensive experience in designing novel Bayesian adaptive clinical trials across all phases of the regulatory process:
- Phase I: Designs to determine maximum tolerated dose levels and other outcomes in safety studies
- Phase II: Preliminary efficacy studies that can accommodate a wide range of data sources, including both clinical and biomarker endpoints
- Phase III: Large-scale pivotal studies that confirm efficacy and safety in a large, heterogeneous population
- Phase IV: Post-marketing studies to monitor adverse event rates and intervene if necessary
Bayesian CID methods enable direct probability statements about treatment effects, including in equivalence studies where we seek to accept (not reject) a null hypothesis of no treatment difference. They can incorporate relevant auxiliary data (including non-randomized, real world data) across a range of applications and data types.
Bayesian methods can also be used ‘seamlessly’ across trial phases, in settings where a tradeoff between safety and efficacy must be made fairly early in the testing process, or where we simply desire to retain early phase subjects and their data for reuse in later phases, enhancing efficiency. PharmaLex statistical staff are well acquainted with CID approaches (and regulator buy-in to these approaches) and auxiliary data incorporation in adaptive trial settings, including the computing needed to simulate design Type I error and power.
Bioinformatics and computational biology play an increasingly important role in the pharmaceutical and biotechnology industries. New biotechnologies are adopted to meet the needs of drug discovery and development. Such technologies produce new and often larger or more complex data types which require either adaption of existing analytical methods, or the development of new methodologies. A thorough understanding of the biological context is also required to enable meaningful interpretation of the results.
PharmaLex’s data scientists develop new strategies to utilize proprietary client data, the vast real-world data and evidence, and reference population databases to support drug discovery and development and other client objectives. Our staff use their multidisciplinary expertise to integrate different data types and sources, use and develop bioinformatics tools, create customized interfaces and visualization tools, and help our clients to find meaningful relationships and results.
We regularly develop and update a variety of analytical pipelines to meet the needs of clinical and non-clinical development and the advent of new biotechnologies. For example:
- Preprocessing and quality control of OMICS data such as genomics, transcriptomics, proteomics, and metabolomics data
- Integration and analysis of vast phenotypic information often seen in Electronic Health Records or biobanks
- Statistical analysis using next generation sequencing data (e.g. whole genome sequencing, bulk RNA sequencing, single-cell or single-nuclei RNA sequencing) in conjunction with clinical data
- Predictive modeling of multi-OMICS molecular or clinical phenotypes using high dimensional data
- Strategies and computational pipelines for drug target identification and validation
For over 20 years, PharmaLex’s team has been serving pharma and biotech clients to help understand and use biomarkers and develop diagnostics. As data types and new technologies evolve, we evaluate, adapt, or develop new statistical methods tailored to analytical needs.
Some examples of our expertise in addressing analytical challenges include:
- How to discover prognostic or predictive biomarkers from high-dimensional transcriptomics data
- Augmenting analysis of proprietary data by incorporating publicly available biomarker data
- Development of optimal analytical strategies, methods, and pipeline for biomarker data based on a new technology
- Designing biological experiments or biomarker-guided clinical trials
- Implementation of automated workflow and visualization of biomarker data processing and analysis
- Qualification and introduction of bioinformatics tools recently developed in academia for industry use
The size, complexity and distributed nature of today’s data means that rigid and centralized architectures and tools break down. Artificial Intelligence (AI) and machine learning (ML) are key components to stay agile and competitive. PharmaLex uses state-of-the-art AI and ML tools and its long experience in data science to solve our specific customer’s problems. Our healthcare expertise combined with our multidisciplinary team allows us to understand our customer’s very questions and envisage top-notch solutions wherever needed.
We develop predictive methodologies to evaluate model quality as the classical metrics may fail at providing the right conclusion. We stay upfront of the technology by using most recent algorithms such as gradient boosted decision trees and deep neural networks. Our IT infrastructure allows us to handle big data in a short time frame.
To ensure compliance, our statistical experts are working closely with our IT and Quality departments. It allows for example the creation of AI-powered software, that is completely compliant with regulatory environment (GAMP5). This accumulated sum of competencies makes us one of the few being able to develop solutions in full accordance with all the (Bio-)Pharmaceutical regulation (ICH-USP-EMA).
Examples of applications:
- Prioritizing genes for drug targets using a variety of evidence sources and annotations
- Prediction of disease diagnosis or relapse after treatment using biomarker data for developing potential companion diagnostics or other diagnostic tests
- Drug response and drug synergy prediction using DNA or RNA data and identifying patient subgroups with improved treatment effect
- Support of clients in presenting machine learning and predictive modeling results to regulatory agencies
- AI driven decision making in analytical method validation and stability studies
- In-line control of bioreactor using Machine learning
- Detection of counterfeit drugs
- Automated analysis of flow cytometry biomarkers
- Deep Learning for automatic epileptic seizures detection in EEG signals
- Automatic quantification of substances from NMR signals or full DAD-chromatogram
- Machine learning problem root cause analysis
- Computer vision enabled diagnostic of pathological tissues to identify disease
“A picture is worth a thousand words”
This adage is particularly true when considering the visual representation of data, especially complex and/or large data. An interactive image can help to see relationships, find trends, or identify the “needle in the haystack”. At PharmaLex, we create dynamic visualization tools that look at data from every angle and use them as a means of communication between the analytical team and the scientist or end-user.
Examples of expertise and experience include:
- Develop exploratory visualization tools to allow for client’s interactive data exploration.
- Design an easy query interface for integrated data from multiple sources (e.g., clinical and molecular).
- Visualize highly complex longitudinal biomarker data (e.g. ctDNA) for custom populations and identifying patterns that inform about patient efficacy
- Generate standalone reproducible reports.
- Create production-level tables, figures, and listings for submission/publication.
Software development
Automation and machine learning are revolutionizing many industries enabling increased productivity, improved quality and enhancing return on investment. The pharmaceutical industry is still largely relying on established, but increasingly outdated, systems and methodologies that were efficient many years ago yet now prove to be a limiting factor in innovation and growth. At PharmaLex we combine our extensive global experience and deep understanding of pharmaceutical lifecycle processes with the development of our cloud-based platform for life sciences automation. Through our pharma technology-enabled services we are helping our customers to adopt and take advantage of innovative technologies to drive efficiency gains within their own operations. For scientists working in CMC development we propose several fully qualified smart solutions such as e.noval and Seelva for the validation of analytical procedures and (relative) potency assays, Transval for supporting transfers of assays and Stab.e.lity for the evaluation of stability of drug product, computing shelf-life, evaluating internal release limits (IRL) and performing accelerated stability studies.
Beyond those software, PharmaLex team also develop and qualified tailored solutions that involve smart analytics and automate analyses and reporting ready for submission.
Applications
Proven Expertise from Discovery to Post Marketing
From drug target discovery to launch and post launch activities, developing a new medical product generates vast amounts of data that need to be processed, analyzed and interpreted.
Whether for a drug, biologic or medical device, PharmaLex statisticians and data scientists have years of experience turning all that data into useful knowledge.
We can help you choose the best strategy for identifying the right drug target and de-risk the decision-making process by using Bayesian statistics to evaluate the probability of success.
Throughout clinical development, our team supports the design, strategy and full analysis of clinical trials. Should you need to include clinical biomarkers, our design and analytical expertise will guide you to success.
Based on our analytical and regulatory proficiency, we also guarantee the seamless development and implementation of any analytical procedures and manufacturing processes.
OUR SERVICES
Drug target identification and validation are critical steps in the drug development process. Identifying the right targets with favorable efficacy and safety profiles can increase the success rate in the more expensive clinical stages of drug development. The increasing affordability of high-throughput omics technologies leads to exponentially growing data thoroughly describing putative targets (e.g., DNA, RNA, protein, metabolite) and providing necessary evidence to qualify and/or validate those targets. At PharmaLex, our multidisciplinary team combines expertise in statistics, medicine, genetics, machine learning computational biology and data visualization to develop new strategies for drug target identification. We have experience with data from many different omics platforms, developing analysis pipelines with thorough quality control checks and interpreting results in the context of all available evidence including clinical data and data from biobanks, knowledgebases, or disease specific consortia.
Contact us to learn more about:
- Strategies on data sources, analysis methods, implementation and interpretation for drug target identification and validation
- Assessment of human genetics evidence of drug targets in existing clinical programs
- Development efficient computational pipelines for large-scale data with state-of-the-art methodologies
- Analysis of data from next generation sequencing (e.g., whole genome sequencing, bulk RNA sequencing, single-cell or single-nuclei RNA sequencing) and other technologies
- Utilizing data from biobanks (e.g., UK Biobank), public databases (e.g., The Cancer Genome Atlas (TCGA)), reference populations, and electronic health records (EHR) to complement proprietary data
- Disease-specific drug target identification and evaluation in different therapeutic areas, incl. neurodegenerative diseases (Alzheimer’s, Multiple sclerosis, Parkinson’s), cardiovascular diseases, autoimmune diseases, endocrine and metabolic diseases, hematology and oncology, NASH and liver diseases, and more
- Using machine-learning based approaches and clinico-genomics data to build predictive models for drug target prioritization
The discovery and pre-clinical development of new drug products is a growing challenge for biopharmaceutical industry for at least three main reasons. First, there are more refined technologies to scrutinize the biology of a new drug product. The more data are collected, the higher the risk of false discovery leading to costly investments. The statistics and data science expertise at Pharmalex allows to extract the relevant information using optimized signal processing methods and reduce the rate of false discovery by means of tailored design of experiments and advanced statistical inference. Second, as regularly mentioned in the scientific literature, pre-clinical studies suffer from an lack of reproducibility. Here again our experience in statistics contributes considerably to improve the reproducibility of conclusions with innovative design of experiments and new analytical strategies. Third, given there are several drug products that are potential candidates, PharmaLex experts propose to de-risk the decision-making process using Bayesian statistics to evaluate the predictive probability of success and supporting optimal portfolio management. Indeed what matters the most is to evaluate the chance for a potential drug product to succeed instead of examining if a past study was successful or not.
Clinical development is the key milestone for any drug on the market. Throughout different studies, the investigational product goes through different testing stages in order to assess it’s safety, identify the potential efficacious dose and confirm this later dose. Various strategies in order to maximize the efficiency of these milestones can be put in place. Statistical theory is at the forefront of these as it allows, for example, to determine the sample size based on appropriate simulations, determine the strategy for adaptive trials, analyze the results of the trials and much more.
In this highly regulated environment PharmaLex has a proven knowledge of expertise on a wide variety of domains. PharmaLex knowledge’s even pushes the boundaries of current regulatory statuses to help in the design of very complex studies. Further, PharmaeLx’s proven experience of Bayesian statistical methods allows to incorporate various sources of uncertainty in order to account for all the available information in the design of clinical trials.
In Brief, PharmaLex has experience in the following domains:
- Design of adaptative clinical trials from very early phase I to phase IV post marketing studies
- Leveraging pre-clinical and PK/PD data to design early phase clinical trials
- Sample size computation
- Innovative Bayesian methods to optimize clinical trial design
- Bayesian Meta-analysis of multiple clinical trials
- Clinical biomarker analysis and interpretation
- Incorporation of uncertainty in the design of the trial design in order to maximize the probability of success of the trial.
- Interactions with different regulatory bodies
- Analyses of clinical trials
- PK/PD analysis
Biomarkers play a crucial role in the era of precision medicine. The Biomarkers, EndpointS and other Tools (BEST) glossary explains a biomarker as a defined characteristic that is measured as an indicator of normal biological processes, pathogenic processes, or responses to an exposure or intervention, including therapeutic interventions. Molecular, histologic, radiographic, or physiologic characteristics are types of biomarkers. Applications of biomarkers are being utilized in basic biomedical research and medical product development.
For over 20 years, we offer biomarker consulting and analytical services to the pharmaceutical and biotechnology industry. PharmaLex has been a pioneer in the field of translational sciences, specializing in biomarker analytics, and has become a preferred service provider of biomarker statistics for major pharmaceutical companies.
Analytical support from discovery through clinical development and application:
- Identification and validation of disease state biomarkers;
- Analysis of pharmacokinetic (PK) markers to determine dosing;
- Linking pharmacodynamic (PD) biomarkers to the mechanism of action (MOA) of a therapeutic intervention;
- Analysis of predictive or prognostic biomarkers;
- Biomarker-driven clinical trial design and experimental design within trials;
- Utilization of biomarkers to identify and stratify patients with improved efficacy and safety profiles;
- Development and validation of diagnostic tools including co-development of companion diagnostics (CDx);
- Analysis and interpretation of anti-drug antibody (ADA) data.
Experienced in analyzing data from a variety of technologies:
- Sequencing technologies: next generation sequencing (NGS; e.g. whole genome, whole exome sequencing, bulk or single-cell RNA-sequencing, circulating tumor DNA sequencing), immuno-sequencing (e.g. immunoSEQ);
- Microarrays for DNA and RNA: whole genome genotyping, candidate genes, methylation arrays, and gene expression arrays;
- Immunoassays: enzyme-linked immunosorbent assay (ELISA) for proteins, and microsphere-based immuno-multiplexing for cytokines and chemokines, and others;
- Other technologies measuring proteins: immunohistochemistry (IHC), mass spectrometry, proprietary platforms (e.g. Olink) and others;
- Immunophenotyping: flow cytometry, mass cytometry (CyToF);
- Imaging: MRI, CT scans and others;
- And many more.
Analytical Procedures and Bioassays are keystones of the development and manufacturing of drug substances and drug products in the bio-pharmaceutical industry. Evolution of the technologies and regulatory framework puts more and more weight in the need in building quality in early stages of the development of these measurement systems, in gaining knowledge and control of their behavior due to the critical role analytical procedures and bioassays plays such as for process characterization, quality control, shelf-life, biosimilarity assessment and so on. For more than 20 years PharmaLex has provided innovative and state to the art statistical support through-out the entire life-cycle of analytical procedures and bioassays: helping scientists design and develop fit-for-purpose analytical procedures and bioassays, provide guidance in their optimization, qualification and validation, as well as during their transfers from one site to another or during the several technical bridging these assays will undergo.
Support for the life cycle of analytical methods and bioassays:
- Implementing analytical Quality by Design (aQbD) methodology
- Development and Optimization using Design of Experiments (DoEs)
- Strategies and methods for parallelism evaluation in parallel line and parallel curve assays
- Advanced signal processing, multiplex assays
- Non-linear modeling
- Reportable value format optimization and justification
- Qualification and validation of analytical procedures and bioassays
- Comparability studies and bridgings
- Evaluation of capability and uncertainty
- Control strategy and transfer
In CMC statistics, GSDS has many years of experience in proposing robust and innovative statistical solutions for the development, validation, and control of processes to manufacture all types of drugs. We help our customers applying the latest guidance on processes, formulation, and assays lifecycle (e.g. FDA process validation, ICH Q1-Q14), from early development and characterization, process performance qualification (PPQ), assay validation, transfer and routine manufacturing. We have the state-of-the-art expertise in Quality by Design and Analytical Quality by Design. In manufacturing sciences, we also support annual review, acceptance criteria definition, release limits and specification computations, and stability studies.
Bayesian thinking is at the heart of our approach. This provides our c-t customers with a crystal-clear understanding of the solutions proposed and its impact on long-term opportunities. Bayesian Statistics allows for improved conclusions from statistical models (uncertainty management, prediction, probabilities of success).
Our team of experts develop tailored statistical and software solutions that can speed up data analysis and reporting to an order of magnitude, while ensuring full compliance to regulatory guidance.
Areas of activities
- FDA’s Process validation
- Powerful Design of Experiments giving flexibility and cost rationalization
- Quality by Design, ICH Q8 Design Space
- Connecting the dots between Analytical development and Process/manufacturing
- Digitalization, ML/AI
- Validated software development for in-house use of advanced statistics not available through commercial software, and/or tailored so results are compliant to regulatory and Quality bodies, and can directly be uploaded in LIMS
- All fields of drug development (Pre-clinical/Discovery/Toxicity studies, Clinical trails,CMC/Process/Product/Assays, Commercial manufacturing)
- All types of drugs (vaccines, biologicals, small molecules, medical device, personalized medicines/ATMP, Orphan and pediatric drugs, etc.)
- Back-office help during FDA interactions with rapid handling of statistical queries
Just like medical products for human use, animal drugs and medical products have to go through rigorous approval processes. In the US, FDA’s Center for Veterinary Medicine (CVM) regulates and approves new drugs through e.g., new animal drug applications (NADA), abbreviated NADA (ANADA), or conditional approval (CNADA) whereas the United States Department of Agriculture – Animal and Plant Health Inspection Service (USDA-APHIS) regulates veterinary diagnostic kits through the Center for Veterinary Biologics (CVB).
PharmaLex statistics and data science team supports the development of veterinary drugs and devices and is familiar with Veterinary ICH (VICH) compliance & standards and using the same highest statistical standards as practiced for the clinical of human drug products. Our team has supported the development of veterinary medical products in different disease areas (e.g., parasite treatment (fleas and heartworm), oncology, feline immunodeficiency virus). Our experiences include adequate Designs of Experiments (DoEs), statistical analyses for Target Animal Safety (TAS) and Target Animal Effectiveness (TAE) studies, analyses of biomarker data, as well as observational studies of aging and frailty and for the development of diagnostic assays.
Available Resources
WEBINARS
RELATED VIDEOS
Global Reach. Local Presence.
Find your Local Contact
