scPhyloX.estimation.haematopoiesis

Classes

LogLike

Functions

cellnumber(t, xx, r, a, b, k, t0, p, r1, b1)	ODE of cell number changes over time
stem_num(i, t, c0, ax, bx, r, k, t0, r1, b1)	Stem cell number calculator
nstem_num(t, x, c0, ax, bx, r, k, t0, r1, b1)
p_xi(gen, T, c0, ax, bx, r, k, t0, r1, b1)	Probability density function of LR distance
my_loglike(theta, data, args)	Likelihood of lr-dist
para_inference_DE(data[, T, c0, sigma, n_iter, ...])	Mutation rate estimation using DE
mcmc_inference(data, para_prior, T, c0, sigma[, draw, ...])	Mutation rate estimation using DE-MCMC

Module Contents

cellnumber(t, xx, r, a, b, k, t0, p, r1, b1)

ODE of cell number changes over time

Args:

t:

time

xx:

[n_stemcell, n_nonstemcell]

stem_num(i, t, c0, ax, bx, r, k, t0, r1, b1)

Stem cell number calculator

Args:

i:

generation

t:

time

c0:

initial cell number

ax,bx,r,d,k,t0:

parameters

return:

float:

Stem cell number in generation i at time t.

nstem_num(t, x, c0, ax, bx, r, k, t0, r1, b1)

p_xi(gen, T, c0, ax, bx, r, k, t0, r1, b1)

Probability density function of LR distance Args:

n_gen:

generation

T:

time

c0:

initial cell number

ax, bx, r, k, t0, r1, b1:

parameters

Return:

np.array:

Probability density of LR distance at time T.

my_loglike(theta, data, args)

Likelihood of lr-dist

Args:

theta:

parameters, (ax, bx, r, k, t0, r1, b1)

data:

Observed lr dist

args:

paramteres, (time, initial_cell_number, prior_sigma)

Return:

float:

Sum of log-likelihood of given lr dist parameters

para_inference_DE(data, T=20, c0=None, sigma=1, n_iter=100, bootstrape=0, verbose='text')

Mutation rate estimation using DE

Args:

data:

lp-dist

n_iter:

Iterations of de estimation

bootstrape:

Weather using bootstrape to accuratly estimate mutation rate, 0 to turn off.

Return:

tuple:

(accepted parameters, loss, de-estimator)

class LogLike(loglike, data, args)

Bases: pytensor.tensor.Op

itypes

otypes

likelihood

data

args

perform(node, inputs, outputs)

mcmc_inference(data, para_prior, T, c0, sigma, draw=1000, tune=1000, chains=9, est_bx=False)

Mutation rate estimation using DE-MCMC

Args:

data:

Observed lp-dist

data_prior:

mean of prior distributions of all parameters

T:

time of phylodynamics eqns

c0:

initial cell numbers

sigma:

variation of loss function

draw:

Number of smaples to draw

tune:

Number of iterations to tune

chain:

number of chains to sample