/repro_compDAG

Primary LanguageR

Modeling Microbial Community Coalescence via Compositional Directed Acyclic Graphical Models

Author Contributions Checklist Form

Data

Abstract

Our real data analysis consists of two datasets.

  • Female Genital Tract 16S RNA Data：microbial 16S RNA data in women's cervix of uterus and vagina from the Human Microbiome Project 2, which can be accessed via the R package HMP2Data, as a part of the multi-omic microbiome study, pregnancy initiative (MOMS-PI).
  • Oral and Stool Microbiome Data：16S rRNA sequences of longitudinal fecal and oral samples from 97 patients with acute myeloid leukemia (AML) receiving induction chemotherapy.

Availability

Yes, the data to replicate our results is available (we have included it with the submission).

Description

The two datasets are provided in the directory /data and are respectively included in the sub-directory /data/realdata_moms_pi and /data/realdata_aml. Please see the README.md in directory /data for detail.

Code

Abstract

The code provided includes the main code to conduct the compDAG method.

Description

  • 02-1coalescence.stan: stan file to conduct posterior sampling for causal direction $y_1 \rightarrow y_2$ ot $y_2 \rightarrow y_1$.
  • 02-2nocausal.stan: stan file to conduct posterior sampling if there is no causal relationship between $y_1$ and $y_2$.
  • 02compDAG.R :R file to conduct compDAG method.
  • 00load_all.E : R file to load all the required packages and functions.
  • 03bQCD.R: R file to conduct bQCD method.
  • main.R: R file to conduct simulation studies in the paper.
  • PARAMS.R R file to read parameter settings from bash file.

Instructions for Use

Reproducibility

All the figures and tables in the article can be reproduced (i.e.Figures 1-2,Tables 1-6 and Table S.1).

We provided the results of intermediate output in the directory /output to avoid having to rerun computationally-intensive steps of the workflow. You can reproduce these intermediate results by following the instructions in README.md in the directory /output.

How to reproduce analyses (e.g., workflow information, makefile, wrapper scripts)

  • Table 1-4 and Table S.2-3 run: “table1-4_S.2-3_code.R”
  • Table 5 run: “table5_code.R”
  • Table 6 run: “table6_code.R”
  • Table S.1 run: “tableS1_code.R”
  • Figure 1-2 and Figure S.1-2 run: “figure1-2_S.1-2_code.R”
  • Figure 3 run: “figure3_code.R”
  • Figure 4 run: “figure4_code.R”