Examples

This section shows how to use tracts on a real dataset. It analyzes a single population from the 1000 Genomes Project, the African Americans from Southwest US (ASW), using trio-phased local ancestry calls. The examples walk through the main steps of an analysis: the input data, the driver files, the demographic models, running the inference and reading the outputs. Below, we provide a description of the data, the preprocessing steps, links to the examples, and the demographic models considered.

Reproducing the 1000 Genomes analyses of the paper

The data used in these examples are not the same as the data analyzed in the tracts paper. Here, the ASW local ancestry calls are trio-phased, whereas the paper uses the population-phased local ancestry calls from the same 1000 Genomes release. The full analysis of the six admixed populations of the 1000 Genomes Project (ACB, ASW, CLM, MXL, PEL and PUR) presented in the paper is available in the folder example/3pops_sexbiased of the GitHub repository. Its README explains how to download the data, run the inference and produce the tables and figures.

Data

Dataset description

Description of the local ancestry datasets used in this section (1000 Genomes Project).

Data description

Data processing

Preprocessing and formatting steps applied to the data.

Data Processing

Example analyses

ASW population

Example analyses for the ASW population.

ASW population

Demographic models

In this section, we consider three demographic models of varying complexity. The corresponding YAML files are specified below.

One pulse model

The following model is written in the file ppp.yaml.

model_name: One pulse
description: Represents a population X founded t generations ago by EUR, NAT, and AFR.
time_units: generations
demes:
   - name: EUR
   - name: AFR
   - name: NAT
   - name: X
ancestors: [EUR, NAT, AFR]
proportions: [REUR, RNAT, RAFR]
start_time: t

Two pulses model

The following model is written in the file ppp_pxx.yaml.

model_name: Two pulses
description: A population X founded t1 generations ago by EUR, AFR, and NAT, then subsequent EUR migration.
time_units: generations
demes:
   - name: EUR
   - name: AFR
   - name: NAT
   - name: X
ancestors: [EUR, NAT, AFR]
proportions: [REUR, RNAT, RAFR]
start_time: t1

pulses:
   - sources: [EUR]
     dest: X
     proportions: [REUR2]
     time: t2

Three pulses model

The following model is written in the file ppp_xxp_pxx.yaml.

model_name: Three pulses
description: A demes model with flexible parameters. Represents a population X founded t1 generations ago by EUR and NAT, then subsequent AFR migration,
then subsequent EUR migration.
time_units: generations
demes:
   - name: EUR
   - name: NAT
   - name: targetpop
ancestors: [EUR, NAT]
proportions: [REUR, RNAT]
start_time: t1

pulses:
   - sources: [AFR]
     dest: targetpop
     proportions: [RAFR]
     time: t2

   - sources: [EUR]
     dest: targetpop
     proportions: [REUR2]
     time: t3