Skip to content

Dominance linkage disequilibrium

21 fully proved

Dominance LD

StructureTwo-locus diploid genotype under random union of gametes#✓fully proved

A diploid two-locus genotype under random union of gametes: the haplotypes (X₁, Y₁) and (X₂, Y₂) are independent and identically distributed, with square-integrable allele codings.

MathlibExt.Biology.StatisticalGenetics.DominanceLD.RandomUnion · 17 lines
structure RandomUnion (P : Measure Ω) where
  /-- The allele coding of haplotype 1 at locus `A`. -/
  x1 : Ω → ℝ
  /-- The allele coding of haplotype 1 at locus `B`. -/
  y1 : Ω → ℝ
  /-- The allele coding of haplotype 2 at locus `A`. -/
  x2 : Ω → ℝ
  /-- The allele coding of haplotype 2 at locus `B`. -/
  y2 : Ω → ℝ
  memLp_x1 : MemLp x1 2 P
  memLp_y1 : MemLp y1 2 P
  memLp_x2 : MemLp x2 2 P
  memLp_y2 : MemLp y2 2 P
  /-- The two haplotypes are independent. -/
  indep : IndepFun (fun ω => (x1 ω, y1 ω)) (fun ω => (x2 ω, y2 ω)) P
  /-- The two haplotypes are identically distributed. -/
  ident : IdentDistrib (fun ω => (x1 ω, y1 ω)) (fun ω => (x2 ω, y2 ω)) P P
Used by 18

Random Union

DefinitionHaplotype variance at the first locus#✓fully proved

The haplotype variance V_A = E (X₁ - p)² at locus A.

noncomputable def varA : ℝ := ∫ ω, (G.x1 ω - G.freqA) ^ 2 ∂P
Uses 2 · Used by 4

DefinitionHaplotype variance at the second locus#✓fully proved

The haplotype variance V_B = E (Y₁ - q)² at locus B.

noncomputable def varB : ℝ := ∫ ω, (G.y1 ω - G.freqB) ^ 2 ∂P
Uses 2 · Used by 4

DefinitionGametic linkage disequilibrium#✓fully proved

The gametic linkage disequilibrium D = E (X₁ - p) (Y₁ - q).

noncomputable def ldD : ℝ := ∫ ω, (G.x1 ω - G.freqA) * (G.y1 ω - G.freqB) ∂P
Uses 3 · Used by 4

DefinitionAdditive coding at the first locus#✓fully proved

The additive coding a_A = (X₁ - p) + (X₂ - p) (centred allele count) at locus A.

noncomputable def addA (ω : Ω) : ℝ := (G.x1 ω - G.freqA) + (G.x2 ω - G.freqA)
Uses 2 · Used by 5

DefinitionAdditive coding at the second locus#✓fully proved

The additive coding a_B = (Y₁ - q) + (Y₂ - q) at locus B.

noncomputable def addB (ω : Ω) : ℝ := (G.y1 ω - G.freqB) + (G.y2 ω - G.freqB)
Uses 2 · Used by 4

DefinitionDominance coding at the first locus#✓fully proved

The dominance coding d_A = (X₁ - p) (X₂ - p) at locus A.

noncomputable def domA (ω : Ω) : ℝ := (G.x1 ω - G.freqA) * (G.x2 ω - G.freqA)
Uses 2 · Used by 3

DefinitionDominance coding at the second locus#✓fully proved

The dominance coding d_B = (Y₁ - q) (Y₂ - q) at locus B.

noncomputable def domB (ω : Ω) : ℝ := (G.y1 ω - G.freqB) * (G.y2 ω - G.freqB)
Uses 2 · Used by 4

TheoremAdditive linkage disequilibrium is twice the gametic one#✓fully proved

Additive LD: E(a_A a_B) = 2 D.

theorem integral_addA_mul_addB : ∫ ω, G.addA ω * G.addB ω ∂P = 2 * G.ldD
Uses 6 · Used by 1

TheoremSecond moment of the additive coding at the first locus#✓fully proved

The additive coding has second moment E(a_A²) = 2 V_A.

theorem integral_addA_sq : ∫ ω, G.addA ω ^ 2 ∂P = 2 * G.varA
Uses 4 · Used by 1

TheoremSecond moment of the additive coding at the second locus#✓fully proved

The additive coding has second moment E(a_B²) = 2 V_B.

theorem integral_addB_sq : ∫ ω, G.addB ω ^ 2 ∂P = 2 * G.varB
Uses 4 · Used by 1

TheoremDominance linkage disequilibrium is the square of the gametic one#✓fully proved

Dominance LD: E(d_A d_B) = D².

theorem integral_domA_mul_domB : ∫ ω, G.domA ω * G.domB ω ∂P = G.ldD ^ 2
Uses 6 · Used by 1

TheoremSecond moment of the dominance coding at the first locus#✓fully proved

The dominance coding has second moment E(d_A²) = V_A².

theorem integral_domA_sq : ∫ ω, G.domA ω ^ 2 ∂P = G.varA ^ 2
Uses 4 · Used by 1

TheoremSecond moment of the dominance coding at the second locus#✓fully proved

The dominance coding has second moment E(d_B²) = V_B².

theorem integral_domB_sq : ∫ ω, G.domB ω ^ 2 ∂P = G.varB ^ 2
Uses 4 · Used by 1

TheoremAdditive and dominance codings at different loci are uncorrelated#✓fully proved

The additive coding at A and the dominance coding at B are uncorrelated: E(a_A d_B) = 0.

theorem integral_addA_mul_domB : ∫ ω, G.addA ω * G.domB ω ∂P = 0
Uses 5

TheoremCorrelation of additive codings equals the gametic correlation#✓fully proved

The correlation of the additive codings is the gametic correlation r = D / √(V_A V_B).

theorem addCorr_eq :
    (∫ ω, G.addA ω * G.addB ω ∂P) /
        Real.sqrt ((∫ ω, G.addA ω ^ 2 ∂P) * ∫ ω, G.addB ω ^ 2 ∂P) =
      G.ldD / Real.sqrt (G.varA * G.varB)
Uses 9 · Used by 1

TheoremDominance LD is the square of additive LD#✓fully proved

Dominance LD is the square of additive LD (Palmer et al. 2023, graphical abstract): the correlation of the dominance codings at two loci is the square of the correlation of their additive codings.

theorem domCorr_eq_addCorr_sq :
    (∫ ω, G.domA ω * G.domB ω ∂P) /
        Real.sqrt ((∫ ω, G.domA ω ^ 2 ∂P) * ∫ ω, G.domB ω ^ 2 ∂P) =
      ((∫ ω, G.addA ω * G.addB ω ∂P) /
        Real.sqrt ((∫ ω, G.addA ω ^ 2 ∂P) * ∫ ω, G.addB ω ^ 2 ∂P)) ^ 2
Uses 14

Dominance LD

DefinitionDominance encoding of a biallelic genotype#✓fully proved

The dominance encoding of a biallelic genotype g ∈ {0, 1, 2} (count of the allele with frequency p): p², -p (1 - p), (1 - p)². Dividing by -p (1 - p) gives the encoding -p / (1 - p), 1, -(1 - p) / p of Palmer et al. (2023).

noncomputable def dominanceEncoding (p g : ℝ) : ℝ :=
  if g = 0 then p ^ 2 else if g = 1 then -(p * (1 - p)) else (1 - p) ^ 2
Used by 1

TheoremDominance coding of allele indicators equals the dominance encoding#✓fully proved

For 0/1 allele indicators the dominance coding (X₁ - p) (X₂ - p) depends only on the genotype X₁ + X₂ and equals its dominance encoding.

theorem domCode_eq_dominanceEncoding {x1 x2 : ℝ} (h1 : x1 = 0 ∨ x1 = 1) (h2 : x2 = 0 ∨ x2 = 1)
    (p : ℝ) : (x1 - p) * (x2 - p) = dominanceEncoding p (x1 + x2)
Uses 1

The dependency explorer needs JavaScript and its graph data.

Download the complete graph data.