Protocol for identifying Mycobacterium tuberculosis infection status through airway microbiome profiling

Abstract

This protocol describes the steps to determine an airway microbiome signature for identifying Mycobacterium tuberculosis infection status. We outline procedures for processing microbiome data, calculating diversity measures, and fitting Dirichlet multinomial mixture models. Additionally, we provide steps for analyzing taxonomic relative and differential abundances, as well as identifying potential biomarkers associated with infection status. For complete details on the use and execution of this protocol, please refer to Kayongo et al.¹.

Keywords: bioinformatics; immunology; microbiology.

Graphical abstract

Figure 1

A combined plot showing microbiome alpha diversity measures according to Mycobacterium tuberculosis Infection Status

Figure 2

A principal component analysis and density plot showing the microbiome beta diversity according to Mycobacterium tuberculosis Infection Status

Figure 3

Dirichlet Multinomial Mixture (DMM) model clusters the airway microbiome into four pulmotypes Adopted from Kayongo et al. with some modifications.

Figure 4

A heatmap showing associations between DMM community types and alpha indices. MetadeconfoundR was used for analysis Cofounded results are shown with a circle, while deconfounded results are shown with a star. Cliff’s delta and FRD values are shown. ∗∗∗ FDR<0.001, ∗∗FDR<0.01 and ∗FDR<0.1.

Figure 5

A PCoA plot showing the microbiome beta diversity among the DMM microbial pulmotypes Adopted from Kayongo et al. with some modifications.

Figure 6

Phylum relative abundance of taxa according to Mycobacterium tuberculosis Infection Status (A) A stacked bar plot showing the most abundant phyla according to Mycobacterium tuberculosis Infection Status. (B) An alluvial plot showing the most abundant phyla according to Mycobacterium tuberculosis Infection Status.

Figure 7

Genus relative abundance of taxa according to Mycobacterium tuberculosis Infection Status (A) A stacked bar plot showing the most abundant genus according to Mycobacterium tuberculosis Infection Status. (B) An alluvial plot showing the most abundant genus according to Mycobacterium tuberculosis Infection Status.

Figure 8

Volcano plots showing enriched and depleted genera according to Mycobacterium tuberculosis Infection Status (A) Shows the enriched and depleted genera between individuals with Active TB and those with Latent TB Infection. (B) Shows the enriched and depleted genera between individuals with Active TB and those that are Uninfected. (C) Shows the enriched and depleted genera between individuals with Latent TB Infection and those that are Uninfected.

Figure 9

A circular tree plot showing taxa with significant LDA scores according to Mycobacterium tuberculosis Infection Status

Figure 10

A bar plot showing phyla with significant LDA scores according to Mycobacterium tuberculosis Infection Status Adopted from Kayongo et al. with some modifications.

Figure 11

Machine Learning Model Performance of rf, SVM and GLM machine learning models (A) Performance of the Random Forest machine learning model, showing an AUC of 99.7. (B) Performance of the Support Vector Machines (SVM) machine learning model, showing an AUC of 59.8%. (C) Performance of the Generalized Linear Model (GLM), showing an AUC of 55.3%.

Figure 12

Significant taxa loaded by the Random Forest machine learning model differentiating Mycobacterium tuberculosis Infection Status Adopted from Kayongo et al. with some modifications.

Figure 13

Cross Validation when running the Random Forest model