Skip to content

Disease progression

These parameters control how individuals move through the SEIR portion of the compartmental model once they have been infected: how fast exposed people become infectious (iota), how new infections split between symptomatic and asymptomatic cases (sigma), how quickly each infectious branch recovers (gamma_1, gamma_2), and how fast natural immunity wanes back to susceptibility (epsilon). Together they set the per-tick transition probabilities used by infectious.py, exposed.py, and recovered.py.

All five are scalars stored as float32, and each is consumed inside the model components via the 1 - exp(-rate) transformation to convert a per-day rate into a per-tick Bernoulli probability for a Binomial draw.

Quick reference

Parameter Shape Dtype Range Off-value
iota scalar float32 >= 0 0
gamma_1 scalar float32 >= 0 0
gamma_2 scalar float32 >= 0 0
epsilon scalar float32 >= 0 0
sigma scalar float32 [0, 1] 1

iota

  • What it controls: Per-day rate at which exposed (E) individuals progress to becoming infectious (I_sym or I_asym).
  • Shape: scalar
  • Dtype: float32
  • Range: >= 0 (validator-enforced; the assert message reads "iota value must be positive" but 0 is allowed).
  • Off-value: 0 — sets the per-tick progression probability 1 - exp(-iota) to zero, freezing the exposed cohort (apart from non-disease deaths) so no new infectious individuals are produced.
  • Consumer code: src/laser/cholera/metapop/exposed.py, src/laser/cholera/metapop/infectious.py
  • Related parameters: sigma, gamma_1, gamma_2
  • Notes: Consumed in infectious.py as _iota_prob = 1 - exp(-iota), then progressing = Binomial(E_next, _iota_prob). The validator wording "must be positive" is slightly inconsistent with the underlying >= 0 assert — 0 passes validation.

gamma_1

  • What it controls: Per-day recovery rate for symptomatic infectious individuals (I_sym -> R).
  • Shape: scalar
  • Dtype: float32
  • Range: >= 0 (validator-enforced; the assert message reads "gamma_1 value must be positive" but 0 is allowed).
  • Off-value: 0 — disables symptomatic recovery; symptomatic infectious individuals never transition to R via this pathway (they can still die through vital-dynamics mortality).
  • Consumer code: src/laser/cholera/metapop/infectious.py
  • Related parameters: gamma_2, iota, sigma
  • Notes: Cached as _gamma_1_prob = 1 - exp(-gamma_1) in infectious.check() and used as the Bernoulli probability for each symptomatic individual's recovery draw. Not listed in the plan §4 inert-values table because disabling recovery is not a feature you would normally toggle off, but it works mathematically.

gamma_2

  • What it controls: Per-day recovery rate for asymptomatic infectious individuals (I_asym -> R).
  • Shape: scalar
  • Dtype: float32
  • Range: >= 0 (validator-enforced; the assert message reads "gamma_2 value must be positive" but 0 is allowed).
  • Off-value: 0 — disables asymptomatic recovery; asymptomatic infectious individuals never transition to R via this pathway.
  • Consumer code: src/laser/cholera/metapop/infectious.py
  • Related parameters: gamma_1, iota, sigma
  • Notes: Cached as _gamma_2_prob = 1 - exp(-gamma_2) and used as the Bernoulli probability for asymptomatic recovery. Not listed in the plan §4 inert-values table.

epsilon

  • What it controls: Per-day rate at which natural immunity wanes (R -> S).
  • Shape: scalar
  • Dtype: float32
  • Range: >= 0 (validator-enforced; the assert message reads "epsilon value must be positive" but 0 is allowed).
  • Off-value: 0 — disables natural-immunity waning; the recovered compartment is monotone (apart from natural mortality) and individuals never return to S via this pathway.
  • Consumer code: src/laser/cholera/metapop/recovered.py
  • Related parameters: omega_1, omega_2
  • Notes: Consumed as _waning_prob = 1 - exp(-epsilon), then waned = Binomial(R - non_disease_deaths, _waning_prob). Matches the plan §4 inert-values row: epsilon = 0 -> R -> S waning disabled.

sigma

  • What it controls: Fraction of new infections that become symptomatic (the remainder are routed to the asymptomatic compartment).
  • Shape: scalar
  • Dtype: float32
  • Range: [0, 1] (validator-enforced; assert message "sigma value must be in the range [0, 1]").
  • Off-value: 1 — routes 100% of new infections to I_sym, leaving I_asym empty; collapses the symptomatic/asymptomatic split to a single branch while preserving the case-reporting pathway (cases are scored against I_sym only). The complementary value sigma = 0 also collapses the split — routing everything to I_asym — but yields zero reported cases even when infections occur, so 1 is the canonical off-value for this docs convention.
  • Consumer code: src/laser/cholera/metapop/infectious.py
  • Related parameters: iota, gamma_1, gamma_2
  • Notes: Used in two places in infectious.py: (1) initial seeding Isym[0] = round(sigma * I_j_initial); Iasym[0] = I_j_initial - Isym[0], and (2) per-tick split of new progressors new_symptomatic = round(sigma * progressing); new_asymptomatic = progressing - new_symptomatic. Either extreme (0 or 1) mathematically collapses the split; choosing 1 as the documented off-value preserves observability through the case-reporting channel.