Module 6

Coat Colour & Pigmentation Biophysics

From melanocyte biochemistry to Turing pattern formation — how genes, diffusion, and thermodynamics paint the cat

1. Coat Colour Genetics

The extraordinary diversity of feline coat colours arises from surprisingly few genetic loci acting on just two pigment types: eumelanin (black/brown) and pheomelanin (red/yellow). The interplay of these loci produces the full palette from jet black to pure white.

The Extension Locus (MC1R)

The melanocortin-1 receptor (MC1R) gene sits at the Extension locus. The dominant allele E permits normal eumelanin production; the recessive allele e impairs MC1R signalling, shifting melanocytes toward pheomelanin synthesis. In cats, this locus interacts with the sex-linked Orange locus in complex ways not seen in most other mammals.

The Orange Locus & X-Inactivation

The Orange (O) locus is unique among mammals — it resides on the X chromosome. The dominant allele O converts all eumelanin to pheomelanin, producing orange cats. The recessive allele o allows normal eumelanin.

The Lyon Hypothesis & Tortoiseshell Cats

Female cats heterozygous at the O locus (X^OX^o) exhibit tortoiseshell or calico patterns. Mary Lyon (1961) demonstrated that in each cell of a female mammal, one X chromosome is randomly inactivated during early embryogenesis. Cells where the X^O chromosome is active produce pheomelanin (orange patches); cells where X^o is active produce eumelanin (black patches). The resulting mosaic pattern is a visible manifestation of X-inactivation at the level of whole organism phenotype.

Since males have only one X chromosome (XY), they are either orange (X^OY) or non-orange (X^oY). Male tortoiseshell cats are extremely rare (~1 in 3,000), typically requiring XXY karyotype (Klinefelter syndrome analogue).

Agouti Signalling Peptide (ASIP)

The Agouti locus controls the temporal switching of pigment type during hair growth. The dominant A allele produces banded (ticked) hairs: alternating eumelanin and pheomelanin bands due to pulsatile ASIP release that periodically antagonises MC1R. The recessive a (non-agouti) allele produces solid-coloured hairs. The agouti pattern is ancestral — it provides camouflage in wild felids.

Hardy-Weinberg Analysis of Coat Colour Alleles

For an autosomal locus with two alleles (e.g., B and b at the Brown locus), the Hardy-Weinberg equilibrium predicts genotype frequencies:

\[p^2 + 2pq + q^2 = 1\]

where $p$ is the frequency of allele B (black) and $q = 1 - p$ is the frequency of allele b (chocolate). If we observe that 4% of cats in a population are chocolate (bb), then:

\[q^2 = 0.04 \implies q = 0.20, \quad p = 0.80\]

\[\text{BB} = 0.64, \quad \text{Bb} = 0.32, \quad \text{bb} = 0.04\]

For the X-linked Orange locus, the equilibrium differs between sexes. Let $p_O$ be the frequency of allele O and $q_o = 1 - p_O$ the frequency of o:

\[\text{Males:} \quad f(\text{orange}) = p_O, \quad f(\text{non-orange}) = q_o\]

\[\text{Females:} \quad f(\text{orange}) = p_O^2, \quad f(\text{tortoiseshell}) = 2p_O q_o, \quad f(\text{non-orange}) = q_o^2\]

In many urban cat populations, $p_O \approx 0.2\text{--}0.3$. With $p_O = 0.25$: approximately 6.25% of females are orange, 37.5% are tortoiseshell, and 56.25% are non-orange, while 25% of males are orange and 75% are non-orange.

2. Turing Pattern Formation

In his landmark 1952 paper “The Chemical Basis of Morphogenesis”, Alan Turing proposed that biological patterns can emerge spontaneously from the interaction of two chemical substances (morphogens) that diffuse at different rates. This reaction-diffusion mechanism provides the mathematical basis for understanding tabby coat patterns in cats.

The Reaction-Diffusion System

Consider two morphogens: an activator $u$ and an inhibitor $v$. They evolve according to:

\[\frac{\partial u}{\partial t} = D_u \nabla^2 u + f(u, v)\]

\[\frac{\partial v}{\partial t} = D_v \nabla^2 v + g(u, v)\]

where $D_u$ and $D_v$ are diffusion coefficients, and $f, g$ describe the local reaction kinetics. The key insight is that the inhibitor must diffuse faster than the activator: $D_v \gg D_u$.

Linear Stability Analysis

To find conditions for spontaneous pattern formation, we linearise around the homogeneous steady state $(u_0, v_0)$ where $f(u_0, v_0) = g(u_0, v_0) = 0$. Setting $u = u_0 + \tilde{u}$ and $v = v_0 + \tilde{v}$, the Jacobian is:

\[J = \begin{pmatrix} f_u & f_v \\ g_u & g_v \end{pmatrix} = \begin{pmatrix} a & -b \\ c & -d \end{pmatrix}\]

where $a = f_u > 0$ (activator self-activation), $b = -f_v > 0$ (inhibitor suppresses activator), $c = g_u > 0$ (activator promotes inhibitor),$d = -g_v > 0$ (inhibitor self-decay). The homogeneous steady state must be stable without diffusion:

\[\text{tr}(J) = a - d < 0 \quad \text{and} \quad \det(J) = -ad + bc > 0\]

With diffusion, the perturbation with wavenumber $k$ grows if the modified determinant becomes negative:

\[\det(J_k) = D_u D_v k^4 - (D_v a - D_u d)k^2 + (bc - ad) < 0\]

The condition for Turing instability requires the discriminant to be positive, yielding the critical diffusion ratio:

\[\frac{D_v}{D_u} > \frac{(b + a)^2}{(b - a)^2}\]

Turing instability condition: the inhibitor must diffuse sufficiently faster than the activator

This means that if we define $d_r = D_v / D_u$, the system transitions from a uniform state to a patterned state when $d_r$ exceeds the critical threshold. The wavelength of the emerging pattern is:

\[\lambda = \frac{2\pi}{k_c} = 2\pi\sqrt{\frac{2D_u D_v}{D_v a - D_u d}}\]

Experimental Validation: Dkk4

For decades, Turing patterns in biology remained theoretical. In 2012, Sick et al. demonstrated reaction-diffusion mechanisms in mouse hair follicle spacing. Then, in a breakthrough study, Kaelin & Barsh (2021) identified the gene Dkk4 (Dickkopf 4) as the activator morphogen responsible for the prepattern in domestic cats. Dkk4 expression establishes the stripe/spot template in cat skin before hair follicle development, directly confirming Turing's 70-year-old mathematical prediction. This was the first definitive identification of the molecular players in a Turing-type pattern in a mammalian coat.

3. Pattern Types & Genetics

The Tabby locus controls the major coat pattern phenotypes in domestic cats. Multiple alleles at this locus, combined with modifier genes, produce the full range of feline patterns.

Mackerel Tabby (T^m/T^m)

Narrow vertical stripes running perpendicular to the spine, resembling a fish skeleton. This is the wild-type pattern, seen in the African wildcat ancestor. Produced by regular spacing of Dkk4 expression bands.

Classic/Blotched Tabby (t^b/t^b)

Wide swirling patterns with a “bullseye” on the flank. The t^b allele is a loss-of-function variant of the transmembrane aminopeptidase Q (Taqpep) gene, which normally constrains stripe width. Recessive to mackerel.

Spotted Tabby

Thought to arise from modifier genes breaking up mackerel stripes into discrete spots. The spotted pattern may represent a 2D Turing instability producing dot-like rather than stripe-like solutions, depending on domain geometry.

Ticked Tabby (T^a dominant)

The dominant T^a allele at the Tabby locus suppresses body stripes while retaining the agouti banding on individual hairs. Seen in Abyssinian and Somali breeds. Stripes remain on legs and tail (incomplete suppression).

White Spotting (KIT Gene)

The White Spotting (S) locus maps to the KIT gene, a receptor tyrosine kinase critical for melanocyte precursor migration from the neural crest during embryogenesis. The S allele causes dose-dependent migration failure:

• ss: No white spotting (complete melanocyte migration)
• Ss: Moderate white spotting (bicolour, “tuxedo”)
• SS: Extensive white (“van” pattern, up to 90% white)

The pattern follows the principle that melanocyte precursors migrate ventrally from the dorsal neural crest. White patches typically appear on the ventral surface (belly, paws, chest) because those are the last regions reached. This is a beautiful example of developmental noise interacting with a genetic threshold — the exact pattern varies between individuals and even between left/right sides of the same cat.

Calico = Tortoiseshell + White Spotting

A calico cat is genetically X^OX^o (heterozygous at the Orange locus) combined with S (white spotting). The white patches created by KIT-deficient melanocyte migration separate the orange and black patches into larger, more distinct domains than in a tortoiseshell (which has the same X-inactivation mosaicism but without white).

4. Siamese Temperature-Sensitive Colouring

The Siamese pointed coat pattern is one of the most elegant examples of temperature-dependent enzyme kinetics in nature. It arises from the c^s allele at the Colour (C) locus, which encodes tyrosinase — the rate-limiting enzyme in melanin biosynthesis.

Molecular Basis

The c^s allele carries a Gly302Arg substitution (glycine to arginine) that makes the tyrosinase enzyme thermolabile. The mutant enzyme folds correctly and functions at lower temperatures but denatures and loses catalytic activity above approximately 33°C. The allelic series at the C locus is:

• C: Full colour (wild-type tyrosinase, stable at all body temperatures)
• c^b: Burmese (less thermolabile, colour reduction less extreme)
• c^s: Siamese (strongly thermolabile, points-only coloration)
• c: Full albino (no tyrosinase activity at any temperature)

Arrhenius Kinetics of Tyrosinase

The temperature dependence of enzyme activity follows the Arrhenius equation:

\[k(T) = A \cdot e^{-E_a / (RT)}\]

where $k$ is the rate constant, $A$ is the pre-exponential factor,$E_a$ is the activation energy, $R = 8.314$ J/(mol·K), and $T$ is absolute temperature. However, for the thermolabile c^s tyrosinase, we must also account for thermal denaturation. The effective activity becomes:

\[v_{\text{eff}}(T) = k(T) \cdot f_{\text{active}}(T)\]

\[f_{\text{active}}(T) = \frac{1}{1 + e^{\Delta H_d (T - T_m) / (R T_m^2)}}\]

where $f_{\text{active}}(T)$ is the fraction of enzyme in the native (active) conformation,$\Delta H_d$ is the denaturation enthalpy, and $T_m$ is the midpoint denaturation temperature. For wild-type tyrosinase, $T_m \approx 50$°C, well above body temperature. For the c^s variant, $T_m \approx 33$°C.

Temperature Map of the Cat

• Body core: ~38.5°C → c^s tyrosinase inactive → no melanin → cream/white
• Ear tips: ~28–30°C → enzyme active → full melanin → dark
• Paw pads: ~29–31°C → enzyme active → melanin → dark
• Tail tip: ~28–30°C → enzyme active → melanin → dark
• Nose: ~31–33°C → partial activity → intermediate colour

This produces the characteristic Siamese “pointed” pattern: dark extremities (points) against a pale body. Importantly, Siamese kittens are born nearly white because the uterine environment is uniformly warm. The points darken as the kitten grows and extremities cool. Similarly, shaving a patch of fur on a Siamese cat's body and keeping it cool will cause the new hair to grow in dark.

Complete effective activity expression:

\[v_{\text{eff}}(T) = A \cdot \exp\!\left(-\frac{E_a}{RT}\right) \cdot \frac{1}{1 + \exp\!\left(\frac{\Delta H_d\,(T - T_m)}{R\,T_m^2}\right)}\]

5. Reaction-Diffusion Pattern Formation

The diagram below illustrates how Turing instability generates spatial patterns. The top panel shows the activator (amber) and inhibitor (blue) concentration profiles evolving from a uniform state. The bottom panel shows the Turing parameter space.

6. Simulations

Turing Pattern Formation & Siamese Tyrosinase Kinetics

This simulation produces two panels: (1) a 2D Turing reaction-diffusion pattern showing tabby-like stripes emerging from random initial conditions, and (2) the Arrhenius/denaturation curve for Siamese tyrosinase activity as a function of temperature.

Python

script.py133 lines

import numpy as np
import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt
import matplotlib.gridspec as gridspec

# ================================================================
# Panel 1: 2D Turing Reaction-Diffusion (Gray-Scott model variant)
# ================================================================
N = 128
Du, Dv = 0.16, 0.08
F, k_gs = 0.035, 0.065
dt = 1.0
n_steps = 8000

# Initialize
u = np.ones((N, N))
v = np.zeros((N, N))
# Seed with random perturbations in center region
np.random.seed(42)
r = 20
cx, cy = N // 2, N // 2
u[cx-r:cx+r, cy-r:cy+r] += 0.05 * np.random.randn(2*r, 2*r)
v[cx-r:cx+r, cy-r:cy+r] += 0.05 * np.random.randn(2*r, 2*r)

def laplacian(Z):
    return (np.roll(Z, 1, 0) + np.roll(Z, -1, 0) +
            np.roll(Z, 1, 1) + np.roll(Z, -1, 1) - 4 * Z)

# Time evolution
for step in range(n_steps):
    Lu = laplacian(u)
    Lv = laplacian(v)
    uvv = u * v * v
    u += dt * (Du * Lu - uvv + F * (1 - u))
    v += dt * (Dv * Lv + uvv - (F + k_gs) * v)

# ================================================================
# Panel 2: Siamese Tyrosinase Arrhenius + Denaturation
# ================================================================
R = 8.314
Ea = 50000.0  # activation energy (J/mol)
A_prefactor = 1e8

T_celsius = np.linspace(15, 45, 300)
T_kelvin = T_celsius + 273.15

# Arrhenius catalytic rate
k_arr = A_prefactor * np.exp(-Ea / (R * T_kelvin))
k_arr_norm = k_arr / np.max(k_arr)

# Denaturation fraction
Tm_wt = 50 + 273.15   # wild-type midpoint
Tm_cs = 33 + 273.15   # Siamese c^s midpoint
Tm_cb = 38 + 273.15   # Burmese c^b midpoint
dH = 200000.0  # denaturation enthalpy (J/mol)

def f_active(T, Tm):
    exponent = dH * (T - Tm) / (R * Tm**2)
    return 1.0 / (1.0 + np.exp(exponent))

f_wt = f_active(T_kelvin, Tm_wt)
f_cs = f_active(T_kelvin, Tm_cs)
f_cb = f_active(T_kelvin, Tm_cb)

# Effective activity
v_wt = k_arr_norm * f_wt
v_cs = k_arr_norm * f_cs
v_cb = k_arr_norm * f_cb

# ================================================================
# Plot
# ================================================================
fig = plt.figure(figsize=(14, 7))
fig.patch.set_facecolor('#0a0a1a')
gs = gridspec.GridSpec(1, 2, width_ratios=[1, 1.2], wspace=0.3)

# Panel 1: Turing pattern
ax1 = fig.add_subplot(gs[0])
ax1.set_facecolor('#0a0a1a')
im = ax1.imshow(v, cmap='inferno', interpolation='bilinear', origin='lower')
ax1.set_title('Turing Pattern (2D Reaction-Diffusion)', color='#fbbf24',
              fontsize=12, fontweight='bold', pad=10)
ax1.set_xlabel('x', color='#9ca3af', fontsize=10)
ax1.set_ylabel('y', color='#9ca3af', fontsize=10)
ax1.tick_params(colors='#6b7280', labelsize=8)
cbar = plt.colorbar(im, ax=ax1, fraction=0.046, pad=0.04)
cbar.ax.tick_params(colors='#6b7280', labelsize=7)
cbar.set_label('Morphogen v concentration', color='#9ca3af', fontsize=8)
ax1.text(N//2, -12, 'Gray-Scott: F=0.035, k=0.065', ha='center',
         color='#9ca3af', fontsize=8)

# Panel 2: Tyrosinase activity
ax2 = fig.add_subplot(gs[1])
ax2.set_facecolor('#0a0a1a')
ax2.plot(T_celsius, v_wt, color='#22c55e', linewidth=2.5, label='Wild-type (C)', alpha=0.9)
ax2.plot(T_celsius, v_cb, color='#60a5fa', linewidth=2.5, label='Burmese ($c^b$)', alpha=0.9)
ax2.plot(T_celsius, v_cs, color='#f59e0b', linewidth=2.5, label='Siamese ($c^s$)', alpha=0.9)

# Body regions
ax2.axvspan(28, 33, alpha=0.15, color='#60a5fa', label='Extremities (28-33 C)')
ax2.axvspan(36, 39, alpha=0.15, color='#ef4444', label='Body core (36-39 C)')
ax2.axvline(x=38.5, color='#ef4444', linestyle='--', alpha=0.5, linewidth=1)
ax2.text(38.7, 0.85, 'Core\n38.5 C', color='#ef4444', fontsize=8)
ax2.axvline(x=30, color='#60a5fa', linestyle='--', alpha=0.5, linewidth=1)
ax2.text(26, 0.75, 'Ears/\npaws', color='#60a5fa', fontsize=8)

ax2.set_title('Tyrosinase Activity vs Temperature', color='#fbbf24',
              fontsize=12, fontweight='bold', pad=10)
ax2.set_xlabel('Temperature (C)', color='#9ca3af', fontsize=10)
ax2.set_ylabel('Effective enzyme activity (normalized)', color='#9ca3af', fontsize=10)
ax2.set_xlim(15, 45)
ax2.set_ylim(0, 1.05)
ax2.legend(fontsize=8, loc='upper left', facecolor='#1a1a2e', edgecolor='#334155',
           labelcolor='#d1d5db')
ax2.tick_params(colors='#6b7280', labelsize=8)
ax2.spines['bottom'].set_color('#334155')
ax2.spines['left'].set_color('#334155')
ax2.spines['top'].set_visible(False)
ax2.spines['right'].set_visible(False)
ax2.grid(True, alpha=0.15, color='#475569')

fig.suptitle('Feline Coat Biophysics: Pattern Formation & Temperature-Sensitive Colouring',
             color='#f59e0b', fontsize=14, fontweight='bold', y=0.98)

plt.savefig('output.png', dpi=150, bbox_inches='tight',
            facecolor='#0a0a1a', edgecolor='none')
print("Simulation complete.")
print("Left: 2D Turing pattern from Gray-Scott reaction-diffusion model.")
print("Right: Tyrosinase activity curves showing Siamese temperature sensitivity.")
print("At body core (38.5 C), Siamese tyrosinase is nearly inactive -> cream fur.")
print("At extremities (28-30 C), enzyme is active -> dark points.")

Click Run to execute the Python code

Code will be executed with Python 3 on the server

References

Turing, A. M. (1952). “The Chemical Basis of Morphogenesis.” Philosophical Transactions of the Royal Society B, 237(641), 37–72.
Kaelin, C. B., McGowan, K. A., & Barsh, G. S. (2021). “Tabby pattern genetics — a whole new breed of cat.” Trends in Genetics, 37(12), 1062–1074.
Kaelin, C. B. et al. (2012). “Specifying and sustaining pigmentation patterns in domestic and wild cats.” Science, 337(6101), 1536–1541.
Lyon, M. F. (1961). “Gene action in the X-chromosome of the mouse (Mus musculus L.).” Nature, 190, 372–373.
Eizirik, E. et al. (2003). “Molecular genetics and evolution of melanism in the cat family.” Current Biology, 13(5), 448–453.
Lyons, L. A. (2015). “DNA mutations of the cat: the good, the bad and the ugly.” Journal of Feline Medicine and Surgery, 17(3), 203–219.
Murray, J. D. (2003). Mathematical Biology II: Spatial Models and Biomedical Applications. Springer, Chapter 2: Turing mechanism for pattern formation.
Sick, S., Reinker, S., Timmer, J., & Schlake, T. (2006). “WNT and DKK determine hair follicle spacing through a reaction-diffusion mechanism.” Science, 314(5804), 1447–1450.
Imes, D. L. et al. (2006). “Albinism in the domestic cat (Felis catus) is associated with a tyrosinase (TYR) mutation.” Animal Genetics, 37(2), 175–178.
David, V. A. et al. (2014). “Endogenous retrovirus insertion in the KIT oncogene determines white and white spotting in domestic cats.” G3: Genes, Genomes, Genetics, 4(10), 1881–1891.

← Module 5: Olfaction & Chemical Sensing Module 7: Neuroscience & Behavior →