Overcoming the plant’s chemical non-homogeneity for accurate potency results
Potency testing cannabis presents unique challenges. Principally, the non- homogeneity of cannabinoids, within the plant, between strains, and even within a single sample, hinders the accurate testing and labeling of products. A potency test of one cannabis flower will not adequately represent other flowers from the same crop or from the same plant.
Current testing practices struggle to overcome the problem of cannabis’chemical non-homogeneity in an affordable, practical way. Industry-standard high-pressure liquid chromatography (HPLC) tests are slow, require significant equipment overhead, and are not appropriate for the needs of cultivators, wholesalers, or consumers.
While chromatography is invaluable and well-justified in many applications, it cannot satisfy the testing needs of a wholesale transaction or assure consumers of product potency at the point of sale. Additionally, chromatography destroys the test sample, creating a Catch-22 for consumer testing needs.
To address cumbersome and destructive testing practices, some cannabis professionals are turning to spectroscopic methods that are far quicker and cheaper. But these technologies are, on their own, questionably accurate. An emerging hybrid method offers a solution to impracticable or otherwise inaccurate testing. By collating multiple spectroscopic tests from every angle of the cannabis flower and refining the technology with digital image analysis researchers have found an accurate method for non-destructive testing. Backed by robust data science and intensive calibration against industry-standard chromatography, the new technology may prove ideal for potency testing cannabis flower.
As legal cannabis expands into a projected $34.1 billion international industry by 2021, [Zhang] active ingredient testing becomes increasingly important. The legitimization of cannabis as a medicine calls for pharmaceutical dosage consistency, and recreational legalization requires labelling similar to alcohol and tobacco products. For cultivators, wholesalers, and others in the supply chain, accurate testing is equally important.
Companies need to know the quality (i.e., chemical composition) of the crop they’re buying or they open themselves to significant risk; sellers need to properly batch their crops to ensure clients receive consistent potency; and consumers want to feel assured of the products they’re consuming. Yet the need to accurately and quickly test cannabis presents myriad challenges.
The Challenge of Non-Homogeneity
Relative to other pharmaceutical and food products, cannabis sativa is hard to potency test. In part, this is because the cannabinoid content of the plant varies widely [Potter].
In cannabis, chemical non-homogeneity occurs:
• between strains and within strains [Royal Seeds]
• between crops of the same strain [Figure 1]
• between individual plants of the same crop [Potter]
• between the flowers from the same plant [Namdar]
• …and even within the divided material of individual flowers [Wilks]
The cause of non-homogeneity is threefold. Through decades of cannabis horticultural, cultivators have bred plants to enhance desirable characteristics and diminish crop vulnerabilities. These manipulations have manifested customized strains exhibiting higher or lower tetrahydrocannabinol (THC) and cannabidiol (CBD) compositions. Secondly, within any species of animal or plant, individuals will have closely-related, yet slightly different, genotypes (defined as the genetic material dictating the range of characteristics that an organism may express). Within that strain-specific range of genotype possibilities, environmental conditions will determine a phenotype (defined as the observable characteristics expressed by an individual plant).
Even plants clonally propagated from the same “mother” plant will exhibit phenotypic traits based on their environmental histories. And within a selection of plants displaying similar phenotypic traits, different chemotypes (defined as the chemical constitution of an individual plant) may occur independently of readily observable characteristics. Microclimates in the grow room or outdoors, inconsistencies in the fertigation system, pests, or other factors can affect plant-to-plant potency.
Finally, the chemical non-homogeneity of cannabis flowers occurs naturally based on their location within the plant structure or even within a single flower. Cultivators commonly observe, and research has documented, more potent flowers at the top of the canopy as compared to those at the bottom. Proximity to the light source plays a causal role in determining the potency of individual flowers [Namdar] so intra-plant potency variance is unavoidable.
Testing Blind: The Problem of Sampling
Industry stakeholders have come to acknowledge a critical problem, one that bypasses even world-class testing
In a large crop of non-homogeneous cannabis, which individual flowers do you select for testing?
Some experts believe that poor sample selection practices and improperly “batched” crops can cause mislabeled potencies varying up to 75% from actual. [T&T Magazine] For health-compromised consumers who rely on cannabis as a medicine, such wild irregularities are clearly unacceptable.
For solutions to the non-homogeneity problem, field-testing should look to how government regulators and laboratories address the issue. Though their elaborate chromatography methods are not feasible for transactional testing, their approach to sampling sheds light on how to mitigate non-homogeneity in the field. Current testing methods estimate an averaged potency with extensive and random crop sampling. Some jurisdictions require random flower selections totaling 0.7% of the overall batch weight. [CA regs] Flowers are selected from the top, middle and bottom of the batch to ensure a representative and random sample. Then, the sampled flowers are ground together, and the mixture is assumed homogenous. But, as sources note, [Sexton] [Rigdon] glandular trichomes, the most potent part of the plant, may fall through the grinding mechanism or settle at the bottom of the mixture. Cannabis testing presents challenges even to HPLC.
HPLC isn’t a feasible solution for field-testing, but the method of combining multiple samples illuminates the solution to flower-to-flower non-homogeneity. Faced with the need to test non homogenous crops, the food industry has turned to quick, spectroscopic testing methods for some applications. But for cannabis, the chemical variations within a single flower warrant a more refined approach.
Near-infrared spectrometry (NIRS) is a spectroscopic form of testing: it uses the light spectrum to assess the chemical contents of the test subject. By beaming particular wavelengths of light onto an object and detecting the wavelength intensities that bounce back, spectrometers estimate the chemical contents of a test specimen without altering it. For testing high-dollar crops, this non-destructive technique is valuable indeed!
NIRS is not as precise as chromatography. For any single test, HPLC is no doubt more accurate. Yet NIRS is appropriate for many applications and approved by the U.S. Food and Drug Administration for medical procedures, [Scheeren] pharmaceutical testing, [Morisseau] and food testing. [Osborne]
Non-homogeneous crop analysis has employed NIRS by averaging the results of multiple samples. And, because NIRS testing takes around 60 seconds (rather than 45 minutes), the “collate and average” approach has worked. For foraging materials like hay, sampling 20 test specimens has allowed farmers to overcome the crop’s non-homogeneity to find an acceptably accurate active ingredient profile. [Putnam]
For NIRS to be viable for a given chemical’s quantitation, spectroscopic engineers must carefully calibrate the equipment for that chemical of interest. Scientists repeatedly correlate the wavelength/intensity results of the spectrometer against gold-standard technologies like HPLC to ensure accurate results. The greater the number of correlations against HPLC, the more robust the NIRS results. [European Medical Agency]
Because cannabinoids are a new test subject for NIRS, the number of correlations and calibrations against HPLC results are scant. Many cannabis-specific NIRS units on the market now are not sufficiently correlated with HPLC and, without an extensive database of cannabinoid specific algorithms, their accuracies suffer.
But the key issue plaguing existing NIRS cannabis technology isn’t an inherent lack of accuracy potential; it’s the way a spectrometer reads its test subject, and, again, the problem stems from cannabis non-homogeneity. Because only a few square millimeters are exposed to the light source during the test processes, and because trichrome distribution on the flower is uneven, current NIRS testers struggle for accuracy.
NIRS and Single-Flower Non-Homogeneity
Extensive research by GemmaCert Ltd. has documented the issue of single-flower non-homogeneity. And, similar to other research into flower non-homogeneity [Wilks], the results show significant variations in potency within the material of a single cannabis flower. To further our understanding of non- homogeneity and potential solutions.
GemmaCert scientists partitioned twenty cannabis flowers into three to six parts, depending on size. Then, each flower partition was potency tested for THC and CBD via industry standard HPLC techniques. Large variations were observed within the flower, with some differences amounting to +/-25% of the averaged potency. This research shows that, because NIRS units test only a small area of a few square millimeters, single- flower potency variance may skew results significantly.
A compromise solution would be to test a single sample multiple times with NIRS to achieve an averaged result. Presumably, this could overcome the potency variations within the sample as in other food industries. Running several tests on different areas of the material would still be significantly faster than HPLC testing and, assuming an adequate library of HPLC-correlated results, improve accuracy to acceptable levels. Yet additional research and development has turned up new techniques. By incorporating supporting technologies, NIRS testing for cannabis can be more accurate, and quicker, while still leaving the test material unaffected.
The Hybrid Testing Solution
The GemmaCert tester is built on a foundation of NIRS testing and the “collate and average” approach. But the proprietary design of the GemmaCert unit uses adjunct technologies to fully realize the benefits of the NIRS, while avoiding the drawbacks. GemmaCert uses NIRS, and motion mechanics. The GemmaCert unit takes multiple measurements, depending on user-preferred settings, reflecting trade-off between accuracy and duration. By testing multiple surfaces on the whole-flower sample, the unit provides a highly-accurate, “collated and averaged” result, yet still keeps test durations in the 1-to-3-minute range. And because the accuracy of NIRS technology is critically dependent on the distance of the test material from the detector, careful manipulation of the detector versus the specimen not only improves accuracy but advances the overall NIRS science.
GemmaCert uses visual analysis
A simplistic examination of a cannabis flower, even without a microscope or magnifier, can reveal uneven trichrome distribution. Because an individual NIRS test accesses only a small surface area, understanding trichrome distribution and flower shape can improve results, even in the case of advanced “collate and average” approach. Advanced digital image analysis helps ensure optimal calibration of the machine.
GemmaCert uses data science and machine learning
NIRS technology is only as good as the quantity and quality of its correlations with HPLC. With over 2,500 flowers correlated with HPLC results, the GemmaCert has far surpassed the data point libraries of other cannabis-specific NIRS testers. That means each of the multiple measurements it performs during a single test has industry-leading accuracy.
The GemmaCert benefits from machine learning too. Cloud-based software analyzes the multitude of results to identify any outlying data. In a continual refinement of accuracy, the software analyzes its own analysis for constant improvement. Then, the test results data is made available to users via
smartphone or laptop.
By combining leading-edge NIRS methodology, visual image analysis, extensive data science, and machine learning, GemmaCert provides a testing solution that’s more than the sum of its parts.
For more information on GemmaCert and how GemmaCert technologies can benefit your company, visit www.gemmacert.com or email firstname.lastname@example.org
Figures to the left depict trichome non- homogeneous distribution:Top depicts the flower Middle identifies trichomes Bottom integrates identified trichomes in a density map
GemmaCert is a biotechnology company, based in Israel since 2015, aiming to become a market leader in medicinal plant composition and potency analysis, starting with cannabis. GemmaCert’s skilled team of chemists, molecular biologists, biotechnologists, data scientists and programmers work tirelessly to advance cannabis analytical solutions. In the long run, GemmaCert’s breakthrough technology will enable patients and doctors to correlate cannabis composition with specific health conditions, significantly enhancing therapeutic treatment by cannabis and transforming the medical cannabis industry.
Postal address: 8 Hamasger Street, P.O.Box 4377, Ra’anana Israel Web: www.gemmacert.com
Trademarks and Copyright GemmaCert is a Trademark of GemmaCert Ltd. Copyright @ 2017 GemmaCert Ltd. All rights reserved
Information in this document is subject to change without notice and does not represent a commitment on GemmaCert Ltd.
GemmaCert Ltd is not liable for errors contained in this document or for incidental or consequential damages in connection with furnishing or use of this material.
GemmaCert products are protected by U.S. and international copyright laws.
Zhang, M. The Global Marijuana Market Will Soon Hit $31.4 Billion But Investors Should Be Cautious. Forbes 2017.
Potter, D.J.; Clark, P.; Brown, M.B.; Potency of D9–THC and Other Cannabinoids in Cannabis in England in 2005: Implications for Psychoactivity and Pharmacology. J Forensic Sci. 2008.
Basic Cannabis Knowledge: Genotype and Phenotype. Royal Queen Seeds. 2016.
Namdar, D.; Mazuz, M.; Ion, A.; Koltai, H.; Variation in the compositions of cannabinoid and terpenoids
in Cannabis sativa derived from inflorescence position along the stem and extraction methods. Industrial
Crops and Products. 2018.
Wilks, D. Testing for Truth Part 2: Is the tip of a single bud more potent than the bottom? Orange
Cannabis Sampling: The Elephant in the Industry. Terpenes and Testing Magazine. Mar/Apr 2018
California Code of Regulations. Title 16 Division 42. Bureau of Marijuana Control. Chapter 5: Testing
Sexton, M; Ziskind, J; Sampling Cannabis for Analytical Purposes. BOTEC Analysis Corp. I-502
Project #430-1e. 2013.
Rigdon, A; Preliminary Protocol for Sample Preparation for Cannabinoid Potency Analysis in Plant and
Edible Matrices. Restek Corporation. 2016.
Scheeren, T.W.; Schober, P.; Schwarte, L.A.; Monitoring tissue oxygenation by near infrared spectroscopy
(NIRS): background and current applications. J. of Clinical Monitoring and Computing. 2012.
Morisseau, K. M.; Rhodes, C. T.; Pharmaceutical Uses of Near-Infrared Spectroscopy. J. of Drug
Development and Industrial Pharmacy. 1995.
Osborne, B.G.; Fearn, T.; Hindle, P.H.; Practical NIR spectroscopy with application in food and beverage
Putnam, D; Recommended Principles for Proper Hay Sampling. National Forage Testing Association.
www.foragetesting.org/index.php?page=exam_info2 Accessed March 2018.
European Medicines Agency. Guideline on the use of Near Infrared Spectroscopy (NIRS) by the
pharmaceutical industry and the data requirements for new submissions and variations. 2012.