Biomarker discovery from patient cells has a well-documented reproducibility problem, and a large share of it originates before any analysis begins. Sample handling, cohort definition and donor variables that were never recorded produce differences that look like biology and survive into publication because nothing in the workflow was designed to catch them.
Pre-analytical variables outrank most analytical ones
The interval between collection and processing changes cell composition and activation state. Time at room temperature changes it further. Whether samples were processed fresh or after freezing changes subset representation, since monocytes recover poorly. Whether cases and controls were handled identically determines whether any difference you find is about disease or about logistics.
That last one is the killer. If patient samples were collected in a clinic and control samples in a research facility, the two groups differ in handling before they differ in anything else, and the resulting signature is real, reproducible within the study, and about the wrong thing.
Cohort definition does more work than sample size
| Variable | Why it has to be captured |
|---|---|
| Confirmed diagnosis | Self-report does not define a cohort |
| Disease activity and stage | Active and quiescent disease give different cells |
| Treatment status | Therapy reshapes the compartment being measured |
| Age and sex | Both shape immune composition independently of disease |
| Cytomegalovirus status | Substantially remodels the T cell compartment |
| HLA genotype | Several conditions carry strong associations that stratify cleanly |
An underpowered study with a tightly defined cohort frequently outperforms a larger one assembled loosely, because the variance it has to overcome is smaller.
Related product
Cryopreserved PBMCs. Single-donor PBMCs in current catalogue formats, with the donor record attached.
Matched components make a finding interpretable
A cellular signature is more convincing when soluble measurements from the same donor point the same way. Serum and plasma collected alongside the cells allow that, and they are straightforward to obtain at collection and frequently impossible to obtain afterwards.
OrganaBio documents matched serum and plasma alongside cellular material for lupus donors, and matched components are generally worth requesting at the point the request is scoped rather than after the design is finished.
Validation is a different sourcing problem
A discovery cohort establishes a candidate. A validation cohort tests it, and it has to be genuinely independent, which means different donors rather than a held-out portion of the same collection.
This is where sourcing planning pays off. A program that secured access to a defined donor population, with the ability to obtain a second independent set later, can validate. A program that bought one batch and analyzed it has a candidate and no route to testing it.
What to specify when sourcing
The indication and the cohort definition including activity and treatment criteria. Matched controls, and how they will be matched. The cellular material and format, with cases and controls in the same format. Matched serum or plasma if soluble measurements are planned. Donor annotation including HLA and viral serostatus. And whether a second independent cohort will be needed for validation.
OrganaBio documents disease-state material across 24 autoimmune indications for research use with a viability specification of greater than 80% post-thaw. Cohort considerations are covered in autoimmune research with disease-state donors and sourcing more broadly in buying human biospecimens.
Frequently asked questions
Why do PBMC biomarker studies often fail to reproduce?
A large share of the problem is pre-analytical. Time to processing, time at room temperature, fresh versus frozen handling and whether cases and controls were handled identically all produce differences that resemble biology and survive into publication.
What is the most damaging design error in biomarker discovery?
Collecting cases and controls under different conditions, such as patients in a clinic and controls in a research facility. The groups then differ in handling before they differ in anything else, and the resulting signature is reproducible within the study and about the wrong thing.
Which cohort variables have to be captured?
Confirmed diagnosis, disease activity and stage, treatment status, age and sex, cytomegalovirus status and HLA genotype where the condition carries an association. A tightly defined smaller cohort frequently outperforms a larger loose one.
Why request matched serum or plasma?
Because a cellular signature is far more convincing when soluble measurements from the same donor point the same way. Matched components are straightforward to obtain at collection and frequently impossible to obtain later.
What makes a validation cohort valid?
Genuine independence, meaning different donors rather than a held-out portion of the same collection. A program that bought one batch has a candidate and no route to testing it.
Does treatment status need to be an inclusion criterion?
Yes. Immunomodulatory therapy reshapes the compartment being measured, so cells from treated patients reflect disease and drug together. Whether that is acceptable depends on the question and should be stated rather than discovered.
What should I specify when sourcing biomarker study material?
Indication and cohort definition including activity and treatment criteria, matched controls and the matching approach, cellular material and format with cases and controls identical, matched serum or plasma if needed, donor annotation, and whether a second independent cohort will be required.
Talk to OrganaBio
Need this material for a specific process?
Formats, vial sizes and donor characterization vary by product, and custom formats are documented where a process needs them. Tell us what your process requires and the scientific team will confirm what can be supplied.
