Sample Preparation and Quality Control
Note: This page is educational. Real laboratory work requires SOPs, biosafety procedures, institutional approvals, and trained supervision.
TL;DR
Sample preparation is the first analysis step. Before PCR, sequencing, proteomics, flow cytometry, or imaging, the biological material has already been filtered by collection, ischemia time, fixation, storage, extraction, dissociation, and QC. Many "computational surprises" are actually sample-prep artifacts.
1. What can happen to a cancer sample
| Sample state | Common downstream use | Key risk |
|---|---|---|
| Fresh tissue | organoids, flow, single-cell, viable assays | ischemia, dissociation bias |
| Fresh frozen | DNA/RNA/protein, spatial, metabolomics | cold-chain failure, freeze-thaw |
| FFPE | pathology, IHC, FISH, targeted DNA/RNA | fragmentation, formalin artifacts |
| Blood plasma | ctDNA, proteins, metabolites | hemolysis, delayed processing |
| Blood cells | germline DNA, immune profiling | anticoagulant and storage effects |
| Bone marrow | hematologic flow/NGS | hemodilution, clotting |
| Effusion/ascites | cytology, organoids, cfDNA | low tumor fraction, inflammation |
2. Pre-analytical variables
The same tumor can produce different data depending on:
- time from excision to preservation
- temperature during transport
- fixative type and fixation duration
- tumor cellularity and necrosis
- stromal and immune admixture
- blood contamination
- extraction method
- freeze-thaw cycles
- batch order
- operator and instrument differences
When data look strange, check the chain of custody before blaming biology.
3. QC by assay family
| Assay family | QC signals |
|---|---|
| DNA sequencing | DNA amount, fragment size, tumor purity, library yield, depth |
| RNA-seq / RT-qPCR | RNA integrity, DV200/RIN, ribosomal content, mapping rate |
| Proteomics | protein yield, digestion efficiency, peptide IDs, missingness |
| Flow/FACS | viability, singlets, compensation, unstained/FMO controls |
| IHC/IF/FISH | tissue preservation, controls, staining batch, scanner QC |
| Organoids/screens | viability, mycoplasma, passage number, growth rate |
4. Tumor purity matters
A bulk tumor sample is a mixture:
- malignant cells
- fibroblasts
- immune cells
- endothelial cells
- necrosis
- normal adjacent tissue
- blood
Low tumor purity can hide mutations, dilute RNA expression, blur methylation signals, and confuse proteomics. High immune infiltration can be biologically meaningful or a confounder depending on the question.
5. Metadata developers should demand
For every sample, try to capture:
| Metadata | Why it matters |
|---|---|
| sample type and anatomic site | biological comparability |
| collection time and preservation time | degradation and ischemia |
| preservation method | FFPE vs frozen vs fresh |
| tumor percentage | variant detection and expression interpretation |
| necrosis percentage | failed extraction and artifacts |
| prior therapy | treatment-induced changes |
| extraction kit/protocol | batch effects |
| QC metrics | reproducibility and filtering |
| batch and operator | hidden confounding |
6. What technologists can build
- Sample lineage graphs from collection to file.
- QC dashboards that combine wet-lab metrics with computational metrics.
- Batch-effect reports before biological interpretation.
- Metadata validators that block analysis when essential sample fields are missing.
- Tumor-purity-aware pipelines for mutation, expression, and methylation analysis.