Additional file 2 of Gut microbiome of helminth-infected indigenous Malaysians is context dependent

Additional file 2: Figure S1. A geographic map showing the locations of each village and the Kuala Lumpur city in Peninsular Malaysia (stars and numbers) together with a table with other information including states, tribes and subtribes. Figure S2. A flow diagram of the total number of subjects (Orang Asli and urban citizens from Kuala Lumpur) involved in both the pre-anthelmintic and post-anthelmintic of this study. Figure S3. A flow diagram summarizing the bioinformatic analysis from raw reads, 1) Quality filtering, remove human reads and adapter (KneadData), taxonomic classification (Kraken2 and Bracken2), 3) K-mer based approach (Sourmash), 4) Estimation of bacterial growth rate (GRiD) to downstream analysis (A–C) such as beta diversity, alpha diversity, effect size estimation and differential abundance, and 5) Functional genes and pathways analysis using HUMAnN v3.0 and its UniRef 50, Pfam, and MetaCyc pathway databases. Figure S4. Difference in the composition of core microbiota between Orang Asli cohort and KL cohort in different taxonomic rank, which include: A Class, B Order, C Family, D Genus, and E Species. Figure S5. Difference in the composition of core microbiota between different geographical location in different taxonomic rank, which include A Family, B Genus, and C Species. Figure S6. A Bar plot of the top 20 species that best predict the difference of the core gut microbiota between Orang Asli (OA) cohort and Kuala Lumpur (KL) cohort using a Random Forest classification model B and C box plots displaying the selected core microbial species that have high variation between Orang Asli (OA) cohort and Kuala Lumpur (KL) cohort based on the Random Forest analysis. The relative abundances of core microbial species between Orang Asli cohort and KL cohort were tested using Wilcoxon rank sum test. B Species with significant higher abundance in Orang Asli cohort than KL cohort, which include (from left to right): HRGM Genome 3145, Gemmiger sp900539695, and Blautia A sp00043661, respectively. C Species with significant higher abundance in KL cohort than the Orang Asli cohort, which include (from left to right): Megamonas funiformis, Phocaeicola plebeius A, and Bacteroides stercoris, respectively. Figure S7. Effects of geographical location on core gut microbiota and the percentage of unmapped reads in the microbiome. A The core gut microbial species showing the largest variation (cut-off 6.0 for the coefficient of variation) between Orang Asli and Kuala Lumpur cohort in Malaysia across 650 samples. Box plots illustrate the percentage of mapped reads in B RefSeq (i.e., Bacteria, protozoa, fungi, viral, archaea) database, and C Unified Human Gastrointestinal Genome (UHGG) database in different geographical locations. Pairwise comparison between each village and the KL cohort was tested using Wilcoxon rank sum test whereas the comparison for all groups was tested using Kruskal-Wallis. Figure S8. Beta diversity of 650 samples [Orang Asli (OA) and Kuala Lumpur (KL) cohort]. Comparison of pairwise beta diversity of all villages to KL cohort, assessed by Jaccard distance based on distance of A nucleotide k-mer sketches (k = 51), B k-mer sketches (k = 31), and C species level. Pairwise comparison between each village against the KL cohort was tested using Wilcoxon rank sum test. Principal Coordinates Analysis (PCoA) of Jaccard distance based on D genus, E k-mer sketches = 21, F k-mer sketches = 31, and G k-mer sketches = 51 in OA and KL cohort. The individuals from different geographical locations were denoted by different colors. Figure S9. Epidemiology data of the Orang Asli (OA) and Kuala Lumpur (KL) cohort. A Distribution of the age group from OA and KL cohort, the OA and KL cohort were denoted by purple and pink color, respectively. B Distribution of the gender from OA and KL cohort, the female and male cohorts were denoted by blue and yellow color, respectively. C The prevalence of different types of helminthiases, which include Trichuris infection, Ascaris infection and hookworm infection, the heavy, moderate and light infection were detonated by different purple color intensity. Figure S10. Beta diversity comparing the gut microbiome between intestinal helminth-infected and uninfected Orang Asli cohorts at species level. The results were visualized using Non-metric multidimensional scaling (NMDS) plot of A Bray-Curtis (ADONIS: p = 0.001, R2 = 0.035; ANOSIM: p = 0.001, R = 0.149) and B Jaccard distance (ADONIS: p = 0.001, R2 = 0.024; ANOSIM: p = 0.001, R = 0.948) and C Principal Coordinates Analysis (PCoA) of Jaccard distance (ADONIS: p = 0.001, R2 = 0.024; ANOSIM: p = 0.001, R = 0.948). The individuals infected and uninfected with intestinal helminths denoted by blue and red color, respectively. Figure S11. Box plots showing alpha diversity of gut microbiome profile at species level using A Shannon diversity and B Simpson diversity index on individuals infected and uninfected with intestinal helminths; C Shannon diversity and D Simpson diversity index on different numbers of intestinal helminth infection; E Shannon diversity and F Simpson diversity index on individuals infected and uninfected with Trichuris sp. infection; and G Shannon diversity and H Simpson diversity index on different villages. The statistical difference between two groups was tested using the Wilcoxon rank sum test whereas more than two groups was tested using Kruskal-Wallis. Figure S12. Box plots showing alpha diversity (i.e., Richness, Shannon and Simpson diversity index) at species level of gut microbiome profile on Orang Asli who are infected and uninfected with intestinal helminths from A Rasau, B Legong, C Judah, D Sepat, and E Bangkong. The comparison of the alpha-diversity index between helminth-infected and noninfected samples is tested using the Wilcoxon rank sum test. Figure S13. Bubble plot shows bacterial species that are differentially abundant between Trichuris infected and uninfected groups in all samples, as well as specific villages based on the output of the A Multivariate Association with Linear Models (MaAsLin2) and B Analysis of Compositions of Microbiomes with Bias Correction (ANCOM-BC). The size of the bubble is negatively proportional to the p-value. The larger the bubble size displaying the lower p-value. Figure S14. Boxplot shows the differences in the abundance of A Lactobacillus gasseri and B Lactobacillus crispatus in helminthic infections (left) and different villages (right). The statistical difference between two groups was tested using the Wilcoxon rank sum test whereas more than two groups (Village) was tested using Kruskal-Wallis. Figure S15. Bar chart shows the changes in the prevalence of different types of helminthic infections in (left), and the prevalence of the number of helminthic infections (right) among the pre-anthelmintic, 21-day, and 42-day post-anthelmintic. Boxplot showing alpha diversity (i.e., Richness, Shannon and Simpson diversity index) of gut microbiome profile at species level on the B Responder, and C Uninfected. The comparison of the alpha-diversity index between helminth-infected and noninfected samples is tested using the Wilcoxon signed-rank test. Figure S16. Beta diversity comparing the gut microbiome at species level between pre-anthelmintic (blue) and post-anthelmintic (red) among the Responders, visualized using Non-metric multidimensional scaling (NMDS) plot of A Jaccard distance (ADONIS: p = 0.001, R2 = 0.014; ANOSIM: p = 0.001, R = 0.072) and B Bray-Curtis (ADONIS: p = 0.001, R2 = 0.020; ANOSIM: p = 0.001, R = 0.072). Principal Coordinates Analysis (PCoA) coordinate plot showing the beta-dispersion of C Jaccard distance and D Bray-Curtis distance based on gut microbiota profile of the Responders, with pre-anthelmintic (red) and post-anthelmintic (black). E PCoA plot showing the beta diversity of Bray-Curtis distances based on gut microbiota profile of Responders. Figure S17. Beta diversity comparing the gut microbiome at species level between pre-anthelminthic (blue) and post-anthelmintic (red) among the Uninfected, visualized using Non-metric multidimensional scaling (NMDS) plot of A Jaccard distance (ADONIS: p = 0.005, R2 = 0.012; ANOSIM: p = 0.001, R = 0.069) and B Bray-Curtis (ADONIS: p = 0.003, R2 = 0.015; ANOSIM: p = 0.001, R = 0.069). Principal Coordinates Analysis (PCoA) coordinate plot showing the beta-dispersion of C Jaccard distance and D Bray-Curtis distance based on gut microbiota profile of the Uninfected, with pre-anthelmintic (red) and post-anthelmintic (black). E PCoA plot showing the beta diversity of Bray-Curtis distances based on gut microbiota profile of the Uninfected. Figure S18. A Bubble plots of the top 10 bacterial species that differentially abundant between pre-anthelmintic and post-anthelmintic in Responders as well as the Uninfected based on the output of the Microbiome Multivariable Association with Linear Models 2 (MaAsLin2). The size of the bubble is negatively proportional to the p-value. The larger the bubble size displaying the lower p-value and, Line plots showing changes to three of the top differentially abundant bacterial species between pre and post treatment samples from B Responders and C Uninfected individuals, with p-values determined by the Wilcoxon signed-rank test. Figure S19. A Heatmap shows the bacterial species that are associated with treatment response while including infection status and village locations as covariates from MaAsLin2 analysis. Blue for positive association and red for negative association. B Boxplots show the differences in the abundance of CAG245 sp900552135 between different village (top), Helminth infections status (middle) and different response group (Uninfected or Responders) (bottom). The statistical difference between two groups was tested using the Wilcoxon rank sum test (cross-sectional) or Wilcoxon signed-rank test (longitudinal) whereas more than two groups (Village) was tested using Kruskal-Wallis test. C Heatmap shows the bacterial species that are associated with drug response while correcting helminth status from MaAsLin2 analysis. Blue for positive association and red for negative association. Figure S20. A Alpha diversity at species level, which visualized using the line plot of the Richness, Shannon and Simpson diversity indices of the Orang Asli (OA) in pre-anthelmintic, 21-day and 42-day post-anthelmintic groups, for Responders (green) and Uninfected (red) individuals. Alpha-diversity index of three timepoints were compared using the Friedman test whereas two timepoints was compared using the Wilcoxon signed-rank test. There are no statistical differences between groups. B Principal coordinates Analysis (PCoA), C Non-metric multidimensional scaling (NMDS), and D Beta-dispersion of Jaccard distance based on gut microbiota profile at species level of the pre-anthelmintic (purple), 21-day (green), and 42-day post-anthelmintic (gold) from the Responders (ADONIS: p = 0.001, R2 = 0.017; ANOSIM: p = 0.001, R = 0.053) (left) and Uninfected (ADONIS: p = 0.219, R2 = 0.011; ANOSIM: p = 0.001, R = 0.052) (right). Figure S21. Growth Rate Index (GRiD) analysis of the gut bacteria in the Orang Asli (OA) cohort. Correlation matrix of the top 20 gut microbial species that correlate with the infection intensity of Trichuris trichiura in A pre-treatment samples and B among Responders. Figure S22. A Heatmap showing the replication the gut microbial species that are associated with intestinal helminth infection among the Responders. The first vertical side bar encodes the intestinal helminth infection while the second side bar indicates the infection intensity of the Trichuris. B Box plots showing the two bacteria (i.e., uncultured Oscilibacter sp. [left] and Phocaeicola vulgatus [right]) that are significantly negatively correlated with the infection intensity of the Trichuris in Responders. The statistical difference between two groups was tested using the Wilcoxon rank sum test. C GRiD score correlation between bacterial species with the infection intensity of Trichuris among the Responders. The bar chart shows the Spearman’s rank correlation coefficient. D Heatmap showing the replication of the gut microbial species that are associated with albendazole treatment among the uninfected. Samples are shown in row by different timepoints (pre-anthelmintic and post-anthelmintic) whereas the rank of the GRiD score of each bacterium is shown in column. Figure S23. Barplot shows all the pathways that most significantly different between Orang Asli and Urban citizen from Kuala Lumpur (by controlling age and gender) (top), Village (with helminth as covariates) (middle), and Albendazole response (which include helminth and village as covariates) (bottom) based on the output of the Microbiome Multivariable Association with Linear Models 2 (MaAsLin2). The length of the bar corresponds to the value of the significant association (can be either positive or negative). Figure S2. Barplot shows all the gene families that most significantly different between Orang Asli and Urban cohort from Kuala Lumpur (by controlling age and gender), Village (with helminth as covariate) (middle), C Albendazole response (which include helminth and village as covariates) (bottom) based on the output of the Microbiome Multivariable Association with Linear Models 2 (MaAsLin2). The length of the bar corresponds to the value of the significant association (can be either positive or negatively associated). Figure S25. Box plot shows the abundance of the top pathways or gene families based on the output of the Microbiome Multivariable Association with Linear Models 2 (MaAsLin2), which include abundance of A Superpathway of L−tryptophan biosynthesis between Orang Asli and Urban cohort from Kuala Lumpur, B Peptidoglycan biosynthesis II (staphylococci) between different villages, C L-glutamate degradation V between pre- and post-albendazole treatment groups, and D Phosphoenolpyruvate carboxylase between pre- and post-albendazole treatment groups. The statistical difference between two groups was tested using the Wilcoxon rank sum test (cross-sectional) or Wilcoxon signed-rank test (longitudinal) whereas more than two groups (Village) was tested using Kruskal-Wallis test. Figure S26. Summary of the methodology from field work, sample collection, shotgun metagenomic sequencing, and data analysis. Figure S27. Flow diagram of the filtering steps before downstream analysis (beta diversity, alpha diversity, and differential abundance). Figure S28. Methodology for the evaluation of microbial growth rate in relation to helminth infection status in both cross-sectional and longitudinal phase using GRiD analysis.