#021 — Pearson's Chi-Square Test of Independence
Statistical Definition
Tests whether two categorical variables are statistically independent by comparing observed contingency cell counts against expected counts.
Bioinformatics Use Case
Testing whether somatic TP53 mutation status (Mutant vs. WT) is associated with tumor histological grade (I, II, III).
Input Variables
2D contingency table of observed integer counts (r x c, expected cell counts >= 5).
Application
Need to prove a genetic mutation segregates with clinical disease stage? Pearson's Chi-Square test quantifies categorical co-occurrence across large patient cohorts in seconds.
#022 — Fisher's Exact Test
Statistical Definition
Computes the exact hypergeometric probability of observing a 2x2 contingency table as extreme or more extreme under fixed marginal totals.
Bioinformatics Use Case
Gene Ontology (GO) pathway over-representation analysis and rare variant burden testing in small clinical cohorts.
Input Variables
2x2 contingency table of non-negative integers [[a, b], [c, d]].
Application
In rare-disease genomics and pathway enrichment, small sample counts break Chi-square approximations. Fisher's Exact Test delivers exact, unassailable p-values regardless of sample size.
#023 — Hypergeometric Enrichment Test
Statistical Definition
Calculates the exact probability of drawing k or more successes in n draws without replacement from a finite population of size N containing K successes.
Bioinformatics Use Case
Functional enrichment testing of differentially expressed genes against MSigDB Hallmark gene sets.
Input Variables
4 integers: N (background genes), K (pathway genes), n (query DE genes), k (overlap).
Application
Turn a raw list of hundreds of differentially expressed genes into actionable biological mechanisms by pinpointing which pathways are significantly over-represented.
#024 — McNemar's Paired Contingency Test
Statistical Definition
Evaluates marginal homogeneity in a 2x2 table of paired binary outcomes, focusing exclusively on discordant pairs.
Bioinformatics Use Case
Comparing diagnostic sensitivity of two competing variant callers or liquid biopsy assays on the exact same patient cohort.
Input Variables
2x2 table of paired binary classifications [[both_pos, test1_only], [test2_only, both_neg]].
Application
Evaluating a new diagnostic assay against the clinical gold standard on the same patients? McNemar's test isolates the discordant cases to prove superior sensitivity.
#025 — Cochran-Mantel-Haenszel (CMH) Test
Statistical Definition
Tests for a consistent association between two binary variables across K stratified 2x2 tables while controlling for a confounding factor.
Bioinformatics Use Case
Testing allele-disease association across multiple multi-ethnic clinical centers or Evolve-and-Resequence replicate populations.
Input Variables
3D array of K stratified 2x2 contingency tables (2 x 2 x K).
Application
Multi-center clinical trials and multi-ancestry GWAS suffer from confounding strata. The CMH test pools evidence across strata to reveal true, unconfounded genetic associations.
#026 — Cochran-Armitage Trend Test
Statistical Definition
Tests for a linear trend in binomial proportions across k ordered levels of a categorical exposure variable.
Bioinformatics Use Case
Genome-Wide Association Study (GWAS) additive genetic model testing disease prevalence across 0, 1, and 2 risk allele copies (AA, Aa, aa).
Input Variables
2 x k contingency table (Cases/Controls by k ordered genotypes/doses).
Application
In GWAS and dose-escalation studies, risk scales with allele dosage. The Cochran-Armitage Trend Test captures additive genetic effects far more powerfully than unordered Chi-square tests.
#027 — G-Test (Log-Likelihood Ratio Contingency Test)
Statistical Definition
Information-theoretic alternative to Pearson's Chi-square test based on Kullback-Leibler divergence / log-likelihood ratios.
Bioinformatics Use Case
Testing codon usage bias or nucleotide transition/transversion frequencies across genomic regions.
Input Variables
Contingency table of observed integer counts.
Application
For sequence composition and large-scale categorical genomics, the G-test offers additive log-likelihood decomposition across nested genomic annotations.
#028 — Exact Binomial Proportion Test
Statistical Definition
Tests whether the observed proportion of successes in n Bernoulli trials significantly differs from a hypothesized probability p0.
Bioinformatics Use Case
Testing for allele-specific expression (ASE) imbalance where maternal vs. paternal read counts deviate from 0.50.
Input Variables
k (observed successes), n (total trials), p0 (null probability).
Application
Detecting imprinted genes or cis-regulatory mutations requires proving RNA reads deviate from a 50/50 allelic split. The Exact Binomial Test gives an exact probability for every heterozygous locus.
#029 — Two-Sample Z-Test for Proportions
Statistical Definition
Compares independent binomial success proportions between two large cohorts using normal approximation.
Bioinformatics Use Case
Comparing objective clinical response rates (ORR) between treatment and placebo arms in a Phase II oncology trial.
Input Variables
k1, n1 (group 1 successes & total); k2, n2 (group 2 successes & total).
Application
When reporting clinical responder rates to stakeholders, the Two-Sample Proportions Test provides an immediate, transparent confidence interval and p-value for response rate lift.
#030 — Hardy-Weinberg Equilibrium (HWE) Exact Test
Statistical Definition
Tests whether observed diploid genotype frequencies (AA, Aa, aa) conform to expected panmictic proportions (p^2, 2pq, q^2).
Bioinformatics Use Case
Quality control filtering of SNP microarrays and whole-genome sequencing VCF variants to remove genotyping artifacts.
Input Variables
3 non-negative integer counts: n_AA, n_Aa, n_aa.
Application
Genotyping errors ruin GWAS results. Running an exact HWE test automatically flags technical SNP calling artifacts before they corrupt downstream disease associations.
#031 — Barnard's Unconditional Exact Test
Statistical Definition
Unconditional exact test for 2x2 contingency tables that maximizes over nuisance parameters, offering higher power than Fisher's Exact Test.
Bioinformatics Use Case
Small-sample clinical responder vs. non-responder comparisons where only total group sizes are fixed.
Input Variables
2x2 contingency table [[a, b], [c, d]].
Application
In small biomarker trials where every patient counts, Barnard's Exact Test recovers statistical power that Fisher's conservative fixed-margin conditioning throws away.
#032 — Cochran's Q Test for Related Proportions
Statistical Definition
Non-parametric extension of McNemar's test evaluating whether k >= 3 matched binary treatments or classifiers yield identical success rates.
Bioinformatics Use Case
Comparing binary variant detection (Detected vs. Missed) across 4 bioinformatics pipelines on the same 50 benchmark truth variants.
Input Variables
Binary matrix of n subjects/variants (rows) by k >= 3 classifiers/treatments (columns).
Application
Evaluating three or more diagnostic algorithms on the exact same sample set? Cochran's Q pinpoints whether detection rates genuinely differ across