#033 — Pearson Linear Correlation Test
Statistical Definition
Measures the strength and direction of a linear relationship between two continuous variables and tests whether r = 0.
Bioinformatics Use Case
Correlating promoter DNA methylation levels with target mRNA expression across TCGA tumor cohorts.
Input Variables
2 continuous numeric vectors (x, y) of equal length.
Application
Prove a direct linear coupling between two continuous molecular readouts—such as drug concentration and pathway inhibition—with a clean correlation coefficient and p-value.
#034 — Spearman Rank Correlation Test
Statistical Definition
Non-parametric rank-based measure assessing any monotonic relationship (linear or non-linear) between two variables.
Bioinformatics Use Case
Correlating skewed RNA-seq TPM expression with non-normal semi-quantitative immunohistochemistry (IHC) pathology scores.
Input Variables
2 continuous or ordinal vectors (x, y) of equal length.
Application
Biological dose-response curves saturate and rarely stay strictly linear. Spearman rank correlation captures monotonic relationships without being fooled by saturation or outliers.
#035 — Kendall's Tau Rank Correlation Test
Statistical Definition
Non-parametric correlation statistic based on the difference between concordant and discordant observation pairs, robust to ties.
Bioinformatics Use Case
Comparing gene rank lists between two replicate CRISPR screens or clinical ordinal staging systems.
Input Variables
2 continuous or ordinal vectors (x, y) of equal length.
Application
When comparing ranked gene lists or clinical staging scores packed with ties, Kendall's Tau offers superior statistical precision and direct probabilistic interpretation.
#036 — Partial Correlation Test
Statistical Definition
Measures the direct linear association between two variables after removing the confounding linear effects of one or more control variables.
Bioinformatics Use Case
Testing co-expression between two transcription factors while controlling for tumor purity or batch RIN quality score.
Input Variables
2 target vectors (x, y) and 1+ confounding covariate vectors (Z).
Application
Apparent gene co-expression is often a mirage driven by tumor purity or sample degradation. Partial correlation strips away confounders to expose true regulatory wiring.
#037 — Point-Biserial Correlation Test
Statistical Definition
Special case of Pearson correlation measuring association between one continuous variable and one naturally dichotomous binary variable.
Bioinformatics Use Case
Quantifying the correlation between continuous drug IC50 sensitivity and binary BRAF V600E mutation status.
Input Variables
1 binary vector (0/1); 1 continuous numeric vector.
Application
Want a single, intuitive effect-size metric linking a binary genetic mutation to a continuous drug response? Point-Biserial correlation delivers an instant, standardized association scores.
#038 — Distance Correlation (dCor) Permutation Test
Statistical Definition
Non-linear dependence metric that equals zero if and only if two random vectors are statistically independent, tested via permutation.
Bioinformatics Use Case
Detecting complex non-monotonic regulatory relationships (e.g., U-shaped or oscillating circadian gene interactions).
Input Variables
2 numeric vectors or matrices (X, Y) of equal sample count n.
Application
Standard correlations completely miss U-shaped or oscillatory biological feedback loops. Distance correlation detects any non-linear dependency between molecular layers.
#039 — Mantel Matrix Correlation Test
Statistical Definition
Permutation test evaluating correlation between two dissimilarity or distance matrices of identical dimensions.
Bioinformatics Use Case
Testing whether gut microbiome Bray-Curtis dissimilarity correlates with host genetic distance or geographic distance.
Input Variables
2 symmetric distance matrices (D1, D2) of size n x n.
Application
Does microbiome composition mirror host genetics or spatial geography? The Mantel test correlates entire pairwise distance matrices to prove macro-ecological alignment.
#040 — Canonical Correlation Analysis (CCA) Wilks' Lambda Test
Statistical Definition
Finds linear combinations of two multivariate datasets that maximize cross-correlation and tests significance via Wilks' Lambda.
Bioinformatics Use Case
Integrating multi-omics blocks—such as correlating 20 plasma cytokines against 30 metabolomic features in the same patients.
Input Variables
2 multivariate numeric matrices (X of shape n x p, Y of shape n x q).
Application
Bridge two complex omics layers—like metabolomics and proteomics—by identifying the exact multi-analyte signatures driving cross-talk between biological systems.
#041 — Fisher's Z-Test for Comparing Two Correlation Coefficients
Statistical Definition
Applies Fisher's r-to-z transformation to test whether the correlation between two variables differs significantly between two independent groups.
Bioinformatics Use Case
Differential co-expression analysis: testing whether Gene A and Gene B lose their regulatory correlation in tumor vs. normal tissue.
Input Variables
r1, n1 (group 1 correlation & size); r2, n2 (group 2 correlation & size).
Application
Cancer rewires cellular networks. Fisher's Z-test proves when two genes switch from tightly coordinated in healthy tissue to completely decoupled in tumors.
#042 — Intraclass Correlation Coefficient (ICC) Test
Statistical Definition
Assesses reliability, reproducibility, and consistency of quantitative measurements made across replicates, batches, or instruments.
Bioinformatics Use Case
Validating technical reproducibility of a clinical qPCR or mass spectrometry assay across three laboratory sites.
Input Variables
Matrix of n biological subjects (rows) by k technical replicates/raters (columns).
Application
Before deploying a biomarker assay into clinical trials, you need proof of technical reproducibility. ICC quantifies assay reliability across batches and instru