AgeTopicModels (version 0.1.0)

simulate_topics: Simulate genetic-disease-topic structure (step 1)

Description

First step of a two-step procedure to simulate a genetic-disease-topic structure. In this step, all genetic effects act on topic 1.

Usage

simulate_topics(
  topic_number,
  num_snp = 100,
  pop_sz = 10000,
  disease2topic = 0,
  v2t = 20,
  snp2t = 0.04,
  snp2d = 0.15,
  liability_thre = 0.8
)

Value

A list of length 3:

• para: Topic parameters.

• genetics_population: Simulated genotype matrix.

• causal_disease: One simulated binary disease caused by loading on topic 1.

Arguments

topic_number

Number of topics to simulate.

num_snp

Total number of SNPs (default 100; must be >= 60).

pop_sz

Number of individuals (default 10,000).

disease2topic

Disease-to-topic effect size (default 0).

v2t

Number of variants affecting topic 1 (0-20; default 20).

snp2t

SNP-to-topic effect size (default 0.04; informed by UKB).

snp2d

SNP-to-disease effect size (default 0.15).

liability_thre

Liability threshold: proportion set to healthy. For example, 0.8 means the top 20% of liability are set to diseased (default 0.8).

Details

Five configurations across three SNP sets:

1. SNP -> disease -> topic: SNP IDs 1-20; disease ID para$D + 1; topic ID 1.

2. SNP * topic -> disease: SNP IDs 41-60; disease ID para$D + 2; topic ID 1.

3. SNP -> topic -> disease; SNP -> disease: SNP IDs 21-(20 + v2t); disease ID para$D + 3; topic ID 1.

4. SNP -> topic -> disease; SNP + SNP^2 -> disease: SNP IDs 21-(20 + v2t); disease ID para$D + 4; topic ID 1.

5. SNP -> topic + topic^2 -> disease; SNP -> disease: SNP IDs 21-(20 + v2t); disease ID para$D + 5; topic ID 1.

See Also

simulate_genetic_disease_from_topic()

Examples

set.seed(1)
disease2topic <- 0
v2t_small <- 20

# Step 1: simulate topics (fast)
rslts <- simulate_topics(
  topic_number = 2, pop_sz = 1000,
  disease2topic = disease2topic, v2t = v2t_small
)
para_sim            <- rslts[[1]]
genetics_population <- rslts[[2]]
causal_disease      <- rslts[[3]]

# Step 2 (optional): generate diseases from topics
reslt_ds <- simulate_genetic_disease_from_topic(
  para_sim, genetics_population, causal_disease,
  disease_number = 20, itr_effect = 1,
  topic2disease = 2, v2t = 20
)
rec_data <- reslt_ds[[1]]