Article | October 11, 2022
Be More Specific: Solving the Specificity Challenge in Metabolomics
Disease processes are complex. Individuals with the same genetic disorder, symptoms, and/or diagnosis can have varied prognosis and response to treatment. The mechanisms underlying disease can be quite different and may be influenced by a range of internal and external exposures experienced over time.
Sensitive and specific biomarkers allow us to stratify patients by these biological differences.
- A sensitive biomarker identifies individuals with a particular condition or response to treatment (e.g. maximizes inclusion of the right patients into a study – a true positive)
- A specific biomarker changes only in relation to a specific condition or specific drug (e.g. individuals with no biomarker change can be excluded from a study – a true negative)
There is however always a trade-off between biomarker sensitivity and specificity that must be balanced.
Elucidating biomarker specificity has been a key challenge in the field of metabolomics. Small molecule biomarkers provide real-time readouts of biological processes occurring at the tissue level and represent an amalgamation of information stemming from genetic influences, environmental exposures, and the biological functions of other organs. They can be produced in a target or diseased tissue and are readily released into the central circulation for non-invasive capture. These advantages, however, may also impact specificity. A small molecule may be generated or affected by a specific disease or a specific therapeutic, or stem from a specific source like an organ or a dietary change, but once in circulation, it can be difficult to decipher its origin.
At Sapient, we understand that addressing the specificity challenge in metabolomics will accelerate adoption of this burgeoning omics approach, and help more sponsors realize the immense value and predictive power that small molecule biomarkers can bring to drug development.
Sapient’s approach: specificity at scale
Providing population-level database validation
Our most powerful resource is our proprietary Human Biology Database, which is comprised of data from hundreds of thousands of biosamples that have been assayed using our rapid liquid chromatography-mass spectrometry (rLC-MS) platform.
When we discover a biomarker of interest in new client samples, we can query it in our database to determine if it is present in populations with the same disease or drug response. The individuals in our database have been clinically followed for an average of 10-30 years, and we have rich phenotypic data linked with the samples. We can evaluate how the metabolite behaves in broad, diverse populations, and begin to disentangle biomarker changes that relate to a specific condition or drug exposure vs. any number of other internal and external exposures over time.
For example, if in one study we find a blood biomarker that changes following treatment for renal disease, we can use the database to validate if the biomarker is present or absent in a different, larger group of individuals with renal disease and ask further questions such as:
- is the biomarker associated with incident development of fast-progressing renal disease?
- does it associate with quantitative measures of kidney function?
- how does the biomarker fluctuate in people over the course of a day, year, and decade?
We can additionally look across thousands of individuals to determine whether the biomarker is present in other disease states or indications. If found to only correlate with renal disease, even at population-scale, then we can have more confidence that it is highly specific. Alternatively, this search could elucidate additional disease states associated with the biomarker for possible indication expansion.
Using genetics to elucidate a biomarker’s origin
Sapient’s database also includes genetics data for a large number of individuals. Using this data, we can perform genome-wide association studies (GWAS) of a given biomarker to determine its associations with both common and rare gene variants. This approach can identify specific biosynthetic processes and transport systems associated with the biomarker that can help localize the biomarker to a specific organ or mechanistic pathway, providing confirmation that the biomarker came from a specific organ such as the liver or kidney. This not only provides insight into the specificity of the biomarker but can help further elucidate the biology of the biomarker in relation to a disease or the pharmacology of a drug.
Confirming specificity across models and samples
A key advantage of small molecule biomarkers is that they are highly conserved across species. Small molecules originated in earliest yeast and bacteria and are among the most ancient of molecules that are present in humans. This provides a unique opportunity to discover and measure biomarkers across model systems, including cellular systems and preclinical species in which experimental design is more easily controlled and manipulated to understand contributing and confounding factors. It also means small molecule biomarkers are typically highly translatable, increasing the likelihood of finding a biomarker in early discovery research that can be applied in later clinical trials.
If a small molecule biomarker is found to change following drug administration in humans, Sapient can reverse translate this marker in a cellular system and/or animal model in which other variables are minimized. This can confirm if the biomarker is in fact changed by the drug, rather than by factors that are difficult to normalize across people such as diet, exercise level, or medication regimen.
The flexibility of our rLC-MS method also allows us to run discovery screenings on diverse sample types, giving us another approach to elucidate the specificity of a biomarker found in blood. We can run the same method on paired tissue samples to confirm that the biomarker is specific to the target organ.
We believe that this specificity at scale will be integral to moving precision medicine forward, allowing us to apply small molecule biomarkers that are both highly sensitive and specific to a disease state, drug, and/or source in the development of therapies aligned with patients and their unique disease biology.