Skip to Content
🧬 BioQuery is in beta. We'd love your feedback!
GuidesUnderstanding Results

Understanding Results

Learn how to interpret your Query Card and use the results in your research.

Query Card Anatomy

Every query generates a Query Card with these components:

1. Natural Language Answer

The top section provides a plain-English interpretation of the findings:

“DDR1 expression is significantly higher in papillary RCC (KIRP) compared to clear cell RCC (KIRC), with a 2.3-fold increase (p < 0.001).”

This summary is designed to be:

  • Accessible to non-experts
  • Accurate and statistically grounded
  • Ready to share with collaborators

2. Interactive Figure

The visualization varies by analysis type:

Analysis TypeFigure Type
Differential ExpressionBoxplot with individual points
Tumor vs NormalGrouped boxplot
Mutation FrequencyBar chart
Survival AnalysisKaplan-Meier curve

Interactivity features:

  • Hover: See exact values for data points
  • Zoom: Click and drag to zoom into regions
  • Pan: Shift+drag to move around
  • Reset: Double-click to reset view

3. Statistical Results

Key statistics are displayed prominently:

All p-values are two-sided unless otherwise noted.

For expression comparisons:

  • P-value: Statistical significance (typically Wilcoxon rank-sum test)
  • Fold Change: Magnitude of difference between groups
  • Sample Sizes: Number of patients in each group
  • Effect Size: Cohen’s d or similar measure

For survival analysis:

  • Log-rank P-value: Significance of survival difference
  • Hazard Ratio: Risk of event in high vs low expression group
  • Median Survival: Median time to event for each group

4. Methods Text

A grant-ready paragraph describing the analysis:

“Gene expression data were obtained from The Cancer Genome Atlas (TCGA). DDR1 mRNA expression (log2 TPM+1) was compared between papillary renal cell carcinoma (KIRP, n=290) and clear cell renal cell carcinoma (KIRC, n=533) using a Wilcoxon rank-sum test. All analyses were performed using BioQuery (bioquery.io).”

Click Copy Methods to copy this text for your paper.

5. SQL Query

For transparency and reproducibility, the exact SQL query is available:

SELECT sample_id, cancer_type, gene_expression FROM tcga.expression WHERE gene_symbol = 'DDR1' AND cancer_type IN ('KIRP', 'KIRC')

Click View SQL to see the full query.

Interpreting P-values

P-values should be interpreted in context. A small p-value with small effect size may not be biologically meaningful.

P-valueInterpretation
p < 0.001Strong evidence against null hypothesis
p < 0.01Moderate evidence
p < 0.05Weak evidence (traditional threshold)
p ≥ 0.05Insufficient evidence

Consider also:

  • Sample size (large N can yield small p-values for trivial effects)
  • Effect size (is the difference biologically meaningful?)
  • Multiple comparisons (if running many queries, adjust significance threshold)

Interpreting Effect Sizes

Fold Change

For expression comparisons, fold change indicates magnitude:

Fold ChangeInterpretation
> 4Very large difference
2-4Large difference
1.5-2Moderate difference
1-1.5Small difference

Hazard Ratio

For survival analysis:

Hazard RatioInterpretation
HR > 2Strong association with worse survival
HR 1.5-2Moderate association
HR 1-1.5Weak association
HR = 1No association
HR < 1Association with better survival

Sample Size Considerations

Results with small sample sizes (n < 20 per group) should be interpreted cautiously.

BioQuery displays sample sizes prominently. Consider:

  • TCGA data is observational - associations may not be causal
  • Some cancer types have limited samples - rare cancers may have n < 50
  • Subgroup analyses reduce effective sample size

Reproducibility

Every Query Card includes:

  1. Timestamp - When the query was run
  2. Data Version - Which TCGA data release
  3. SQL Query - Exact database query
  4. Statistical Methods - Tests and parameters used

This ensures your results can be reproduced and verified.