Association between human genomes and mycobacterial genomes contributes to severity and transmissibility of pulmonary tuberculosis

Pornpen Tantivitayakul; Punna Kunhapan; Manon Boonbangyang; Kridsadakorn Chaichoompu; Alisa Wilantho; Waritta Sawaengdee; Wuthiwat Ruangchai; Yuttapong Thawornwattana; Hideki Yanai; Sukanya Wattanapokayakit; Areeya Disratthakit; Reiko Miyahara; Supalert Nedsuwan; Worarat Imasanguan; Pacharee Kantipong; Boonchai Chaiyasirinroje; Jody Phelan; Taane G Clark; Martin L Hibberd; Katsushi Tokunaga; Surakameth Mahasirimongkol; Prasit Palittapongarnpim

doi:10.1016/j.isci.2026.117699

. 2026 Sep 22;29(10):117699. doi: 10.1016/j.isci.2026.117699

Association between human genomes and mycobacterial genomes contributes to severity and transmissibility of pulmonary tuberculosis

Pornpen Tantivitayakul ^1,⁴, Punna Kunhapan ², Manon Boonbangyang ⁴, Kridsadakorn Chaichoompu ³, Alisa Wilantho ³, Waritta Sawaengdee ², Wuthiwat Ruangchai ⁴, Yuttapong Thawornwattana ⁴, Hideki Yanai ⁵, Sukanya Wattanapokayakit ², Areeya Disratthakit ², Reiko Miyahara ⁶, Supalert Nedsuwan ⁷, Worarat Imasanguan ⁷, Pacharee Kantipong ⁷, Boonchai Chaiyasirinroje ⁸, Jody Phelan ⁹, Taane G Clark ⁹, Martin L Hibberd ⁹, Katsushi Tokunaga ⁶, Surakameth Mahasirimongkol ², Prasit Palittapongarnpim ^4,^10,^∗

PMCID: PMC13657231 PMID: 42866265

Summary

Pulmonary tuberculosis (PTB) remains a major infectious disease globally. This study aimed to identify determinants of Mycobacterium tuberculosis (MTB) transmission clustering, treatment outcome, and pulmonary cavitation by simultaneously analyzing host and pathogen genomic data. Whole-genome sequencing of 1,960 MTB isolates identified four major lineages (L1–L4). Host population stratification by principal component analysis of genome-wide single nucleotide polymorphisms (SNPs) classified 899 patients (70%) into a genetically homogeneous main group for genome-wide association study (GWAS). Multivariate regression demonstrated that MTB lineage was the primary determinant of transmission clustering, rifampicin resistance was the dominant predictor of unfavorable treatment outcome, and host genetic group was independently associated with pulmonary cavitation. GWAS identified novel loci in collectin subfamily member 10 (COLEC10), CTNNA3, and long intergenic non-coding RNA 2775 (LINC02775) associated with clustering, treatment outcome, and cavitation, respectively, with the association of COLEC10 also found in the remaining group of patients. These findings highlight the importance of integrating bacterial and human genomic analyses to understand tuberculosis host-pathogen interaction.

Keywords: pulmonary tuberculosis, human genetics, MTB lineages, whole genome sequencing, genotyping, transmissibility, unfavorable outcome, lung cavitation

Graphical abstract

Highlights

•
MTB lineage 2 (Beijing) is the main driver of tuberculosis transmission clustering
•
Host and pathogen genetics make distinct contributions to treatment outcome
•
Pulmonary cavitation is primarily determined by host genetics rather than MTB lineage
•
COLEC10 variants implicate collectin-lectin pathway in MTB transmission susceptibility

Genetics; Genomics; Genotyping; Disease in systems toxicology

Introduction

Mycobacterium tuberculosis (MTB) is one of the most remarkable pathogenic bacteria. In 2024, approximately 10.7 million individuals globally developed tuberculosis (TB) and 1.23 million died from the disease.¹ Although MTB can infect almost any organ, pulmonary TB (PTB) is the predominant manifestation of TB in adults and the only significant transmissible form of the disease.

Whole-genome sequencing (WGS) has recently been used to investigate the genetic diversity of MTB. As a clonal organism, MTB is classified into phylogenetically distinct lineages and sublineages according to single nucleotide variant (SNV) phylogeny. MTB sublineages are associated with various factors such as age, geography, ethnicity, and drug resistance.²^,³ Additional associations have been reported with HIV seropositivity, extrapulmonary forms of the disease and their incidence can change over time. Intriguingly, MTB sublineages have rarely been associated with a particular pulmonary form of the disease, such as cavitation.⁴

Genetic investigations of TB patients have been conducted to identify single nucleotide polymorphisms (SNPs) associated with susceptibility to or progression of TB disease.⁵ The studies identified mutations that compromise immune response associated with TB susceptibility, particularly mutations in interleukin 12 (IL-12)-interferon gamma (IFN-γ) axis,⁶ including TNF polymorphism⁷ and variations in the tyrosine kinase 2 (TYK2) gene,⁸ which contributes to the impairment of IFN-γ production.⁹ Nonetheless, conventional genome-wide association studies (GWASs) frequently identify high-effect loci that were not replicable across different populations.¹⁰ This limited replicability likely reflects the small effect size of individual loci, population-specific differences in allele frequencies and linkage disequilibrium (LD) patterns, confounding environmental factors, the phenotypic heterogeneity of TB disease, and the effects of long-term host-pathogen co-evolution.¹¹ Evidence for co-evolution is provided by observed correlation between particular MTB sublineages and specific human genetic susceptibility loci.¹² Other studies demonstrated significant associations between some human SNPs and particular MTB sublineages,¹³ illustrating the interplay between MTB and the human genome in determining the susceptibility to TB infection.

We have conducted a population-based cohort study of TB patients in Chiang Rai, the northernmost province in Thailand, bordering Myanmar and Lao, since the 1990s. Our previous WGS analysis of MTB isolates collected from 2003 to 2010 (designated as HAP1 samples) demonstrated considerable diversity of MTB.¹⁴ The lineages and sublineages were highly correlated with ethnicity, age, geography, and drug resistance. The host GWAS study identified association between certain host SNPs and susceptibility to specific sublineages of MTB.¹² This study aimed to investigate the human SNPs linked to distinct TB phenotypes, specifically the presence of pulmonary cavitation, treatment outcomes, and the MTB transmission clustering. In addition to the HAP1 samples, we included a second set of samples collected from 2015 to 2020 (designated as HAP2), enabling a broader and more comprehensive analysis. The integrated analysis revealed that the phenotypes were mostly associated with both bacterial and host genetic factors.

Results

Demographic and clinical characteristics of TB patients

This population-based study enrolled newly diagnosed, culture-positive TB patients in Chiang Rai province across two periods of time: 1,170 patients from 2003 to 2010 (HAP1) and 790 patients from 2015 to April 2020 (HAP2). The demographic and clinical profiles of all TB patients were shown in Table S1 and described in detail in Data S1.

MTB genotypic diversity and transmission clustering

Among the 1,960 MTB isolates, WGS identified 97,109 SNVs relative to the H37Rv genome. The phylogenetic tree constructed from of all isolates demonstrated high genotypic diversity (Figure S1) and the distribution of genotypes was not significantly different between HAP1 and HAP2 (Table S1).

The patients in Chiang Rai were infected with MTB lineages 1–4 (L1 to L4) with frequencies of 43% (L1), 42% (L2), 1% (L3), and 14% (L4). The frequencies of each lineage and sublineage and their clinical profiles were shown in Tables 1 and S2, respectively. Consistent with previously reported findings for HAP1,¹⁴ and now confirmed for HAP2, MTB lineages were significantly correlated with patient age, ethnicity, anti-TB drug resistance, and bacterial genetic clustering (Tables 1 and S3). Notably, L1 was associated with unfavorable treatment outcome in HAP1, although the association did not reach statistical significance in HAP2. There were no statistically significant differences between the outcomes of various L1 sublineages. MTB lineages were not significantly associated with sex, HIV infection, pulmonary cavitation, or sputum acid-fast bacilli (AFB) positivity (Table 1).

Table 1.

Profiles of patients infected with various MTB lineages

Characteristics	MTB isolates (n = 1,960)
Characteristics	L1	L2	L3	L4	Chi-square, p value
Gender

Male	631 (75%)	591 (71%)	15 (83%)	176 (67%)	χ2 = 7.57, p = 0.056
Female	215 (25%)	239 (29%)	3 (17%)	87 (33%)	χ2 = 7.57, p = 0.056

Age

≤19 years	9 (1%)	36 (4%)	0	9 (3%)	χ2 = 105.86, p < 0.001
20–29 years	61 (7%)	132 (16%)	4 (22%)	37 (14%)
30–39 years	131 (16%)	156 (19%)	4 (22%)	51 (20%)
40–49 years	151 (18%)	171 (21%)	3 (17%)	67 (26%)
50–59 years	190 (23%)	149 (18%)	5 (28%)	45 (17%)
60–69 years	165 (19%)	114 (14%)	1 (5.5%)	37 (14%)
≥70 years	139 (16%)	70 (8%)	1 (5.5%)	15 (6%)

Ethnicity

Thai (%)	676 (88%)	470 (62%)	6 (35%)	114 (47%)	χ2 = 223.65, p < 0.001
Hill tribe and other (%)	88 (12%)	294 (38%)	11 (65%)	128 (53%)	χ2 = 223.65, p < 0.001

HIV status

Negative	731 (88%)	745 (91%)	16 (94%)	238 (92%)	χ2 = 4.46, p = 0.216
Positive	99 (12%)	77 (9%)	1 (6%)	22 (8%)	χ2 = 4.46, p = 0.216

Lung cavitation

Absence	500 (76%)	474 (75%)	10 (67%)	153 (73%)	χ2 = 1.12, p = 0.773
Presence	160 (24%)	159 (25%)	5 (33%)	56 (27%)	χ2 = 1.12, p = 0.773

AFB positivity

Negative	90 (11%)	75 (9%)	2 (11%)	19 (7%)	χ2 = 3.36, p = 0.339
Positive	740 (89%)	748 (91%)	16 (89%)	242 (93%)	χ2 = 3.36, p = 0.339

Grade of AFB positivity

Low (1+ and scanty)	343 (42%)	342 (42%)	6 (35%)	110 (43%)	χ2 = 0.408, p = 0.939
Positive (2+ and 3+)	480 (58%)	468 (58%)	11 (65%)	148 (57%)	χ2 = 0.408, p = 0.939

Treatment outcome

Favorable outcome	626 (85%)	652 (90%)	13 (93%)	214 (91%)	χ2 = 12.07, p = 0.006
Unfavorable outcome	115 (15%)	71 (10%)	1 (7%)	21 (9%)	χ2 = 12.07, p = 0.006

Isoniazid resistance

Sensitive	721 (90%)	680 (87%)	9 (53%)	224 (88%)	χ2 = 25.68, p < 0.001
Resistance	77 (10%)	106 (13%)	8 (47%)	31 (12%)	χ2 = 25.68, p < 0.001

Rifampicin resistance

Sensitive	785 (98%)	756 (95%)	15 (88%)	250 (97%)	χ2 = 13.51, p = 0.004
Resistance	18 (2%)	42 (5%)	2 (12%)	8 (3%)	χ2 = 13.51, p = 0.004

Bacterial genetic clustering

Non-clustered	776 (92%)	527 (64%)	14 (78%)	188 (71%)	χ2 = 229.71, p < 0.001
Small cluster size	64 (7%)	166 (20%)	4 (22%)	61 (23%)
Large cluster size	8 (1%)	137 (16%)	0	15 (6%)

Open in a new tab

Bold entries in Table 1 indicate statistical significance (p < 0.05).

To assess transmission linkage, pairwise genetic distances between all isolate pairs were calculated. A total of 455 isolates (23%) were within 12 SNVs of at least one other isolate. These isolates were clustered into 124 clusters, with sizes ranging from 2 to 33 isolates. Of these, 113 clusters (91%) were small (2–5 isolates), while 11 large clusters comprised 160 isolates (8% of all samples). Strikingly, 8 of the 11 large clusters belonged to MTB L2.2 (the Beijing family) (Figure S2; Table S4).

Human host population structure

Whole-genome human SNP chip typing was performed on 1,293 patients for whom matched genomic data from both M. tuberculosis isolates and human hosts were available. Following quality control and imputation, 5,075,782 SNP positions with minor allele frequency (MAF) > 2% were retained for further analysis.

We performed population stratification analyses using iterative pruning principal component analysis (PCA) (iterative pruning principal component analysis [IPCAPS]), which partitioned the 1,293 patients into 72 PCA nodes. 14 of these were terminal nodes containing more than 10 members (Figure 1A). The total number of patients within these 14 terminal nodes was 1,196; the remaining 97 patients were considered as outliers. The pairwise fixation indices were calculated and shown in Figure S3.

Population genetic structure and ancestry composition of tuberculosis patients revealed by IPCAPS-based clustering and ADMIXTURE analysis

(A) The PCA plots of the first two principal components (PCs) showing population genetic structures of TB patients. Each dot represents a patient and is colored according to its PCA groups.

(B) The ancestor proportions of the 14 identified IPCAPS nodes were estimated by ADMIXTURE. The ADMIXTURE analysis was performed for K (no. of ancestors) = 4, represented as yellow, purple, green, and blue. Each individual is represented by a thin vertical line that is partitioned into color-coded ADMIXTURE proportions reflecting genetic contributions from ancestry.

ADMIXTURE analysis of the 14 terminal nodes identified optimal four to five ancestral population components (Figure S4). PCA nodes 27, 29, and 32 exhibited the lowest pairwise fixation indices (Fst ≤ 0.002, Figure S3) and highly similar ancestry profiles, as they predominantly composed of two major ancestral components (blue and purple) and two minor components (yellow and green) (Figure 1B). Consequently, the 899 patients (70% of total) belonging to PCA nodes 27, 29, and 32 were grouped together into the main group, which was used for GWAS. The remaining 394 patients (30%) constituted the remaining group, in which significant SNPs from the main group GWAS were subsequently tested by candidate gene association analysis to identify possibly reproducible SNP association.

The main group was predominantly composed of patients who self-identified as Thai (93%), while the remaining group comprised members of various ethnic minority communities, most frequently Akha and Lahu, with Thais as the third most common ethnicity (Table S5). The profiles of two patient groups differed significantly in age, ethnicity, prevalence of pulmonary cavitation, treatment outcome, MTB lineage distribution, and the level of bacterial genetic clustering (Table 2). Patients in the main group were older and had a higher proportion of unfavorable treatment outcomes, but were less likely to have pulmonary cavitation and had lower levels of bacterial genetic clustering relative to the remaining group.

Table 2.

Comparisons of the profiles of TB patients belonging to the main group and the remaining group

Characteristics	PCA groups of TB patients		Chi-square, P-value
Characteristics	Main group n = 899	Remaining group n = 394	Chi-square, P-value
Gender

Male	669 (74.4%)	276 (70.1%)	χ2 = 2.65, p = 0.103
Female	230 (25.6%)	118 (29.9%)	χ2 = 2.65, p = 0.103

Age

≤19 years	14 (1.6%)	20 (5.1%)	χ2 = 91.48, p < 0.001
20–29 years	65 (7.2%)	73 (18.5%)
30–39 years	107 (11.9%)	65 (16.5%)
40–49 years	187 (20.8%)	90 (22.8%)
50–59 years	199 (22.1%)	82 (20.8%)
60–69 years	191 (21.2%)	48 (12.2%)
≥70 years	136 (15.1%)	16 (4.1%)

Ethnicity

Thai	842 (93.9%)	57 (14.9%)	χ2 = 799.57, p < 0.001
Non-Thai and hill tribe	55 (6.1%)	325 (85.1%)	χ2 = 799.57, p < 0.001

HIV status

Negative	876 (98.4%)	389 (99.5%)	χ2 = 2.48, p = 0.115
Positive	14 (1.6%)	2 (0.5%)	χ2 = 2.48, p = 0.115

Lung cavitation

Absence	588 (74.1%)	222 (67.7%)	χ2 = 4.69, p = 0.030
Presence	206 (25.9%)	106 (32.3%)	χ2 = 4.69, p = 0.030

AFB positivity

Negative	92 (10.3%)	30 (7.7%)	χ2 = 2.10, p = 0.147
Positive	804 (89.7%)	360 (92.3%)	χ2 = 2.10, p = 0.147

Grade of AFB positivity

Low (1+ and scanty)	278 (35.2%)	128 (36.0%)	χ2 = 0.06, p = 0.813
High (2+ and 3+)	511 (64.8%)	228 (64.0%)	χ2 = 0.06, p = 0.813

Treatment outcome

Favorable outcome	709 (88.3%)	308 (93.6%)	χ2 = 7.25, p = 0.007
Unfavorable outcome	94 (11.7%)	21 (6.4%)	χ2 = 7.25, p = 0.007

NAT2 acetylator phenotype

Slow	186 (29.4%)	107 (35.0%)	χ2 = 3.34, p = 0.189
Intermediate	329 (52.1%)	142 (46.4%)
Rapid	117 (18.5%)	57 (18.6%)

Isoniazid resistance

Sensitive	735 (87.3%)	318 (87.4%)	χ2 = 0.001, p = 0.973
Resistance	107 (12.7%)	46 (12.6%)	χ2 = 0.001, p = 0.973

Rifampicin resistance

Sensitive	817 (95.7%)	356 (97.0%)	χ2 = 1.21, p = 0.271
Resistance	37 (4.3%)	11 (3.0%)	χ2 = 1.21, p = 0.271

MTB lineages

L1	468 (52.1%)	69 (17.5%)	χ2 = 145.14, p < 0.001
L2	339 (37.7%)	226 (57.4%)
L3	9 (1.0%)	6 (1.5%)
L4	83 (9.2%)	93 (23.6%)

Bacterial genetic cluster

Non-cluster isolates	738 (82.1%)	275 (69.8%)	χ2 = 26.96, p < 0.001
Small cluster isolates	99 (11.0%)	78 (19.8%)
Large cluster isolates	62 (6.9%)	41 (10.4%)

Open in a new tab

Bold entries in Table 2 indicate statistical significance (p < 0.05).

Risk factors for MTB clustering, treatment outcome, and pulmonary cavitation

MTB genetic clustering

Multivariate logistic regression identified the period of sample collection and MTB lineage as independent correlates of genetic clustering (Table 3). Infection with L2 (adjusted odds ratio [aOR] = 4.80; 95% confidence interval [CI]: 3.12–7.40) and L4 (aOR = 3.25; 95% CI: 1.87–5.66) were each associated with significantly higher odds of clustering compared to L1. For large-cluster analysis only, MTB lineage remained the sole independent variable, with aORs of 15.85 (95% CI: 5.59–44.94) for L2 and 5.39 (95% CI: 1.50–19.43) for L4 relative to L1 (Table 3). Overall, 36% of patients infected with L2 isolates were in a cluster.

Table 3.

Factors associated with MTB genetic clustering cases and MTB large cluster cases analyzed with univariate analysis and multivariate logistic regression model

Characteristics	MTB genetic clustering cases				MTB large genetic clustering cases
	Univariate analysis		Multivariate analysis		Univariate analysis		Multivariate analysis
	OR (95% CI)	p value	aOR (95% CI)	p value	OR (95% CI)	p value	aOR (95% CI)	p value
HAP

HAP1	1	–	1	–	1	–	1	–
HAP2	0.32 (0.23–0.43)	<0.001	0.41 (0.29–0.58)	<0.001	0.45 (0.29–0.71)	0.001	0.46(0.22–0.96)	0.037

Age

<50 years	1	–	1	–	1	–	1	–
≥50 years	0.42 (0.32–0.55)	<0.001	0.73 (0.51–1.02)	0.068	0.39 (0.26–0.60)	< 0.001	0.64 (0.37–1.12)	0.117

MTB lineages

Lineage 1	1	–	1	–	1	–	1	–
Lineage 2	5.92 (4.13–8.49)	<0.001	4.80 (3.12–7.40)	<0.001	19.63 (7.91–48.74)	<0.001	15.85 (5.59–44.94)	<0.001
Lineage 3	2.95 (0.80–10.85)	0.104	1.78 (0.35–9.19)	0.491	0.00	0.999	0.00	0.999
Lineage 4	4.17 (2.63–6.61)	<0.001	3.25 (1.87–5.66)	<0.001	6.41 (2.16–19.02)	0.001	5.39 (1.50–19.43)	0.010

Host genetic group

Main group	1	–	1	–	1	–	1	–
Remaining group	1.98 (1.51–2.61)	<0.001	1.31 (0.92–1.88)	0.135	1.57 (1.04–2.37)	0.033	0.71 (0.41–1.20)	0.201

NAT2 acetylator phenotype

Slow	1	–	–	–	1	–	1	–
Intermediate	0.95 (0.68–1.33)	0.776	–	–	0.63 (0.39–1.03)	0.067	0.66 (0.38–1.14)	0.139
Rapid	0.74 (0.48–1.16)	0.195	–	–	0.62 (0.32–1.20)	0.154	0.64 (0.30–1.36)	0.242

AFB smear

Negative	1	–	1	–	1	–	–	–
Positive	2.66 (1.45–4.91)	0.002	1.31 (0.67–2.58)	0.429	1.72 (0.73–4.01)	0.210	–	–

Grade of AFB smear

Negative and Low (1+)	1	–	–	–	1	–	–	–
High (2+ and 3+)	0.93 (0.71–1.22)	0.602	–	–	1.06 (0.70–1.61)	0.790	–	–

HIV status

Negative	1	–	1	–	1	–	1	–
Positive	2.22 (0.80–6.17)	0.125	2.41 (0.68–8.63)	0.175	4.04 (1.27–12.68)	0.018	3.25 (0.85–12.51)	0.086

Lung cavitation

Absence	1	–	1	–	1	–	–	–
Presence	0.73 (0.52–1.02)	0.066	0.72 (0.49–1.06)	0.094	0.81 (0.48–1.37)	0.434	–	–

Treatment outcome

Favorable outcome	1	–	–	–	1	–	–	–
Unfavorable outcome	1.038 (0.65–1.67)	0.877	–	–	0.67 (0.28 –1.58)	0.362	–	–

Isoniazid

Sensitive	1	–	1	–	1	–	1	–
Resistance	1.76 (1.21–2.56)	0.003	1.13 (0.66–1.94)	0.648	1.64 (0.96–2.83)	0.073	1.39 (0.73–2.62)	0.317

Rifampicin

Sensitive	1	–	1	–	1	–	–	–
Resistance	2.78 (1.54–5.02)	0.001	2.11 (0.87–5.10)	0.098	1.70 (0.70–4.10)	0.239	–	–

Open in a new tab

Bold entries in Table 3 indicate statistical significance (p < 0.05).

Treatment outcome

The multivariate regression analysis revealed that study period, patient age, patient genetic group, HIV positivity, and rifampicin resistance were independently associated with unfavorable treatment outcome (Table 4). Rifampicin resistance conferred the highest independent risk (aOR = 14.5; 95% CI: 3.55–59.27). Patients in the remaining group had significantly better outcomes than those in the main group (aOR = 0.51; 95% CI: 0.26–0.99). Importantly, MTB lineage, isoniazid resistance, and NAT2 acetylator phenotypes were not identified as independent contributors to outcome in the multivariate model, even though L1-infected patients showed poorer outcomes in univariate analysis (Table 1).

Table 4.

Univariate and multivariate regression analysis to assess factors associated with unfavorable outcome (treatment failure and death) of TB patients

Characteristics	Univariate analysis		Multivariate analysis
Characteristics	OR (95% CI)	p value	aOR (95% CI)	p value
HAP9

HAP1	1	–	1	–
HAP2	0.07 (0.04–0.14)	< 0.001	0.06 (0.03–0.13)	< 0.001

Age

<50 years	1	–	1	–
≥50 years	1.52 (1.02–2.26)	0.039	1.86 (1.07–3.23)	0.027

MTB lineages

Lineage 1	1	–	1	–
Lineage 2	0.56 (0.37–0.86)	0.007	0.67 (0.38–1.19)	0.169
Lineage 3	0.67 (0.08–5.29)	0.701	1.47 (0.12–18.00)	0.763
Lineage 4	0.66 (0.36–1.22)	0.185	1.03 (0.46–2.30)	0.950

Host genetic group

Main group	1	–	1	–
Remaining group	0.51 (0.32–0.84)	0.008	0.51 (0.26–0.99)	0.048

NAT2 acetylator phenotype

Slow	1	–	–	–
Intermediate	1.13 (0.70–1.81)	0.621	–	–
Rapid	1.23 (0.68–2.20)	0.500	–	–

AFB smear

Negative	1	–	1	–
Positive	7.23 (1.76–29.66)	0.006	1.47 (0.26–8.23)	0.658

Grade of AFB smear

Negative and low (1+)	1	–	1	–
High (2+ and 3+)	1.54 (1.01–2.35)	0.045	1.56 (0.90–2.69)	0.111

Lung cavitation

Absence	1	–	1	–
Presence	0.63 (0.37–1.06)	0.080	0.77 (0.43–1.40)	0.394

HIV status

Negative	1	–	1	–
Positive	5.27 (1.73–16.01)	0.003	4.37 (1.10–17.28)	0.036

Isoniazid

Sensitive	1	–	1	–
Resistance	1.55 (0.90–2.68)	0.112	0.99 (0.42–2.38)	0.992

Rifampicin

Sensitive	1	–	1	–
Resistance	4.24 (1.88–9.55)	< 0.001	14.50 (3.55–59.27)	< 0.001

Bacterial genetic clustering

Non-cluster isolates	1	–	–	–
Cluster isolates	1.04 (0.65–1.67)	0.877	–	–

Open in a new tab

Bold entries in Table 4 indicate statistical significance (p < 0.05).

Pulmonary cavitation

Among all covariates examined, only patient genetic group was independently associated with pulmonary cavitation; patients in the remaining group had higher odds of cavitation than those in the main group (aOR = 1.45; 95% CI: 1.06–1.98) (Table 5). The absence of bacterial lineage as a significant factor suggests the possibility of human genetic influence in determining cavitation susceptibility.

Table 5.

Variables associated with lung cavitation in TB patients analyzed with univariate analysis and multivariate logistic regression model

Characteristics	Univariate analysis		Multivariate analysis
Characteristics	OR (95% CI)	p value	aOR (95% CI)	p value
HAP

HAP1	1	–	1	–
HAP2	1.43 (1.10–1.86)	0.008	1.27 (0.93–1.74)	0.138

Age

<50 years	1	–	–	–
≥50 years	1.01 (0.78–1.31)	0.960	–	–

MTB lineages

Lineage 1	1	–	–	–
Lineage 2	1.11 (0.84–1.47)	0.470	–	–
Lineage 3	1.09 (0.34–3.53)	0.889	–	–
Lineage 4	1.00 (0.67–1.51)	0.994	–	–

Host genetic group

Main group	1	–	1	–
Remaining group	1.36 (1.03–1.80)	0.031	1.45 (1.06–1.98)	0.021

NAT2 acetylator phenotype

Slow	1	–	–	–
Intermediate	0.89 (0.62–1.28)	0.541	–	–
Rapid	0.78 (0.48–1.25)	0.301	–	–

AFB smear

Negative	1	–	–	–
Positive	1.31 (0.84–2.04)	0.241	–	–

Grade of AFB smear

Negative and low (1+)	1	–	1	–
High (2+ and 3+)	1.38 (1.05–1.81)	0.020	1.32 (0.98–1.77)	0.066

Treatment outcome

Favorable outcome	1	–	1	–
Unfavorable outcome	0.63 (0.37–1.06)	0.080	0.78 (0.43–1.40)	0.396

HIV status

Negative	1	–	–	–
Positive	0.71 (0.20–2.55)	0.594	–	–

Isoniazid

Sensitive	1	–	1	–
Resistance	0.71 (0.45–1.10)	0.129	0.81 (0.50–1.32)	0.401

Rifampicin

Sensitive	1	–	–	–
Resistance	0.78 (0.35–1.73)	0.535

Bacterial genetic clustering

Non-cluster isolates	1	–	1	–
Cluster isolates	0.73 (0.52–1.02)	0.066	0.76 (0.51–1.12)	0.165

Open in a new tab

Bold entries in Table 5 indicate statistical significance (p < 0.05).

GWAS of host genetic loci in relation to MTB phenotypes

To identify human genetic variants associated with MTB transmission clustering, treatment outcome, and pulmonary cavitation, GWAS was performed exclusively in the genetically homogeneous main group (n = 899). Because the infecting MTB lineage might have influence on the association between host genotype and clinical phenotype, GWAS was conducted both for the entire main group and separately for patients infected with the two most prevalent lineages: L1 (52% of patients) and L2 (38%), which provided a considerable sample size in each group (Table S6). The GWAS revealed several SNPs associated with the phenotypes in the main group. Some of the association may be specific to this population of patients. In order to determine whether the SNPs might be associated with the same phenotypes in other human populations, significant SNP loci identified in the main group GWAS were subsequently tested in the heterogeneous remaining group (n = 394) by candidate gene association analysis.

Genome-wide significant SNPs across the main group

GWAS across the entire main group identified eight SNPs that passed the genome-wide significance threshold. Three intronic SNPs in collectin subfamily member 10 (COLEC10) were associated with increased MTB genetic clustering, with rs1485302 G>A yielding the highest OR (2.13; 95% CI: 1.66–2.72). One SNP, rs7079284 T>C in an intron of CTNNA3, was highly significantly associated with unfavorable treatment outcome (p = 6.34 × 10⁻¹⁰; OR = 0.38; 95% CI: 0.28–0.53). One SNP, rs4655484 C>T in long intergenic non-coding RNA 2775 (LINC02775), was significantly associated with pulmonary cavitation (OR = 2.78; 95% CI: 1.90–4.06). Manhattan plots for clustering, outcome, and cavitation are shown in Figures 2 and S5–S8. A summary of all significant SNPs is provided in Table 6.

Genome-wide association analysis identifying host genetic variants associated with *Mycobacterium tuberculosis* transmission clustering

Manhattan plot and quantile-quantile (Q-Q) plots show the results from GWAS between 161 patients infected with clustered MTB isolates and 738 patients with non-clustering MTB. The x axis represents the genome in physical order; the y axis shows −log₁₀(p value) for all SNPs. The dotted blue line and red line denote the threshold at 1 × 10⁻⁵ and 5 × 10⁻⁸, respectively. The Q-Q plots represent correlation between expected and observed p values.

Table 6.

Human genetic loci associated with unfavorable outcomes, lung cavitation, and genetic clustering of MTB

Phenotype	rs ID^a	CHR	Position	REF>ALT	Host locus	Host annotation	MTB lineages	AF (%)	No. of SNPs^b	OR (95% CI)	p value
Clustering of TB cases	rs1485302	8	118963905	G>A	COLEC10	gene (intron)	All lineages	45.27%	3	2.13 (1.66–2.72)	1.91E-08
Unfavorable outcome	rs7079284	10	66669805	T>C	CTNNA3	gene (intron)	All lineages	78.83%	1	2.59 (1.88–3.58)	6.34E-10
	rs4886299	13	61050306	A>C	LINC00378	gene (intron)	L1	19.69%	1	3.08 (2.03–4.69)	3.23E-08
	rs74896300	3	89226931	G>A	EPHA3	gene (intron)	L2	3.64%	1	7.88 (3.20–19.40)	3.70E-08
	rs372684410	19	35739123	C>T	IGFLR1	gene (intron)	L2	2.81%	1	9.98 (3.68–27.05)	2.23E-08
Lung cavity	rs4655484	1	214184108	C>T	LINC02755	gene (intron)	All lineages	7.49%	1	2.78 (1.90–4.06)	1.94E-08

Open in a new tab

REF, nucleotide allele is on reference genome; ALT represents the alternate allele.

OR, odds ratio; 95% CI represents 95% confidence interval.

AF (%) refers to the percentage of allele frequency of ALT allele.

Bold entries in Table 6 indicate statistical significance (p < 0.05).

the SNP with the strongest association (minimum p value).

Number of SNPs with p < 5 × 10⁻⁸.

Lineage-stratified GWAS

Three additional SNPs were significantly associated with the unfavorable outcome in either L1-infected patients or L2-infected person as shown in Table 6 and Figure 3. Among L1-infected patients, rs4886299 A>C in LINC00378 was significantly associated with unfavorable treatment outcome. Among L2-infected patients, two SNPs were significant: rs74896300 G>A in ephrin type-A receptor 3 (EPHA3) and rs372684410 C>T in insulin growth factor-like family receptor 1 (IGFLR1) (Table 6; Figure 3). In all cases, the allele frequencies of the relevant SNPs were comparable between L1- and L2-infected subgroups, indicating that the differences in GWAS results were not attributable to the frequencies of the alleles.

Lineage-specific associations between host genetic variants and unfavorable treatment outcome, comparing *Mycobacterium tuberculosis* lineage 1 and lineage 2 infections

Log-transformed odds ratios (ORs) with 95% confidence intervals are presented for the association between three SNPs; *rs4886299* A>C (red circles), *rs74896300* G>A (purple squares), *rs372684410* C>T (blue triangles), and the risk of unfavorable outcome stratified by MTB lineages (L1 and L2) based on the GWAS of the main group patients. Each point represents the estimated log(OR) for unfavorable outcome, with error bars indicating 95% confidence intervals. The horizontal dashed line at log(OR) = 0 denotes no association.

Candidate gene association in the remaining group

The three COLEC10 SNPs associated with clustering in the main group were replicated in the remaining group (p = 0.016; OR = 1.47; 95% CI: 1.08–2.00), remaining significant after Bonferroni correction (threshold p < 0.017). The COLEC10 A allele frequency was 45.27% in the main group and 52.16% in the remaining group. None of the SNPs associated with treatment outcome or pulmonary cavitation in the main group were replicated in the remaining group (Table S7), likely reflecting the greater genetic heterogeneity and smaller sample size of the remaining group, and possibly reflecting population-specific effects.

Discussion

It is widely accepted that MTB strains vary in virulence; however, the precise mechanisms underlying these differences remain poorly understood. Virulence can be broadly categorized into two distinct but interrelated dimensions, infectivity and disease severity. In this study, the parameter most closely associated with infectivity was MTB genetic clustering, which reflects both the transmission rate and the probability that infected individuals will subsequently develop active TB disease.¹⁵^,¹⁶ In contrast, treatment outcomes are partly related to disease severity, if the effect of drug resistance is excluded. Pulmonary cavitation is indirectly associated with both dimensions of virulence, as it reflects substantial lung tissue destruction, is usually accompanied by sputum positivity, and is presumed to enhance the likelihood of transmission.¹⁷^,¹⁸

The role of host genetic factors in determining these three phenotypic outcomes has rarely been investigated, particularly in the context of concurrent bacterial genetic contributions. Several previous studies revealed the interactions between human genetic variations and MTB sublineages in determining TB susceptibility.¹¹ In the present study, we extended this approach by systematically evaluating the contributions of both bacterial genotypes and host genetic groups to all three phenotypes, while considering for potential confounding factors.

MTB genetic clustering reflects epidemiological linkage and may therefore be associated with the transmissibility of MTB. Multivariate logistic regression analysis in the main group identified MTB lineage as the principal independent determinant of genetic clustering, with L2 and L4 isolates being more frequently clustered than L1 isolates, suggesting greater transmission potential. This is consistent with the observation that 8 of the 11 large clusters belonged to L2, and aligns with findings from several previous studies.¹⁹^,²⁰^,²¹ Although host genetic group was not identified as an independent predictor of clustering, GWAS analysis of the main group identified three SNPs in the COLEC10 gene associated with clustered cases. This association was subsequently confirmed in the remaining group. Collectin-10 (CL-L1), encoded by COLEC10, shares structural similarity with mannose-binding lectin (MBL) and functions in the innate immune system by binding carbohydrates on the bacterial cell wall. CL-L1 forms a heterodimer with collectin-11 (CL-K1), known as CL-LK, which binds mannose-capped lipoarabinomannan on the mycobacterial cell envelope. Serum CL-LK levels have been reported to be reduced in active TB patients compared to healthy controls.²² These data suggest that COLEC10 variants may modulate susceptibility to MTB transmission, although functional validation is needed.

Regarding treatment outcomes, several socioeconomic factors are known to influence TB treatment success. From a biological perspective, our findings are consistent with previous studies²³ indicating that rifampicin resistance is the strongest independent predictor of unfavorable treatment outcome. This supports the importance of rapid rifampicin resistance diagnostics in TB control programs. Although isoniazid mono-resistance has been implicated in treatment failure in some studies,²⁴^,²⁵ it was not identified as an independent factor in the present analysis, possibly because isoniazid resistance is frequently co-occurring with rifampicin resistance in clinical settings.

Multivariate analysis further demonstrated that MTB lineages were not independently associated with treatment outcome, while patient age and host genetic groups were significant independent predictors. Remarkably, patients in the main group, who predominantly self-identified as ethnic Thais, had approximately twice the odds of an unfavorable outcome compared to the remaining group, most of whom self-identified as hill-tribe individuals. This difference is unlikely to be explained by differential access to healthcare facilities alone.

Studies of host genetic aspects affecting TB treatment outcome have focused predominantly on the pharmacogenetics of anti-tuberculous drugs.²⁶ However, other genetic factors also contribute to treatment outcome. Notably, even before the advent of anti-tuberculous therapy, advanced TB was not invariably fatal,²⁷ suggesting that host genetic factors influence survival. In the present study, GWAS of the main group identified rs7079284 T>C in CTNNA3 as significantly associated with the unfavorable outcome (p = 6.34 × 10⁻¹⁰). CTNNA3 encodes αT-catenin, a key structural protein involved in cell-cell adhesion and cellular integrity.²⁸ Variants in CTNNA3 have been implicated in several diseases, including toluene diisocyanate-induced asthma, although their roles in TB pathogenesis or treatment response have not previously been explored.²⁹

Few studies have reported associations between specific MTB sublineages and treatment outcomes,³⁰^,³¹ but MTB sublineage was not identified as an independent variable in the present analysis. Mokrousov et al. previously reported that ancestral Beijing isolates with intact RD181 (equivalent to L2.2.AA1) were associated with high mortality and reduced pulmonary cavitation.³⁰ In contrast, sublineage L2.2.AA1 was rare in the present study (n = 17, approximately 1%) and was not associated with elevated mortality. This sublineage is more prevalent in other countries, where it has been linked to outbreaks and extensive drug resistance.³² These observations suggest that virulence may vary among clades within the same sublineage, or that host-pathogen interactions differ across geographic regions.

Pulmonary cavitation is a common manifestation of TB, reported in 29%–87% of TB patients,³³ and has been associated with increased transmission risk and prolonged treatment duration.¹⁸ In the present study, cavitation was not associated with MTB lineage, consistent with several previous reports.⁴^,³⁴ There is limited evidence supporting a direct role of intrinsic MTB molecules in mediating tissue destruction and contributing to lung cavitation.³⁵^,³⁶ Current models propose that MTB indirectly induces complex immune responses, leading to granuloma formation and finally cavitation. Our finding that host genetic group was an independent factor associated with pulmonary cavitation in the main group supports the hypothesis that host genetic variation, and its corresponding influence on immune responses, plays a central role in determining cavitation.³⁷ The SNP rs4655484 C>T is located within LINC02775 on chromosome 1q32.3. More than 8,000 large intergenic noncoding RNAs (lincRNAs) have been described and some lincRNA-associated SNPs have been implicated in TB susceptibility and disease progression,³⁸^,³⁹ the functional role of LINC02775 and rs4655484 in TB remains unstudied and warrants further investigation. The absence of an association between MTB genotype and cavitation does not exclude a potential role for MTB genetic factors in cavitation pathogenesis. A recent study identified several homoplastic variants in Rv2828, rpoC, and rpoB that were independently associated with radiological severity.⁴⁰ These variants are not genotype-specific; thus, they did not contradict the results here. This may also explain the variations in pulmonary pathology caused by different MTB strains in animal models.⁴¹^,⁴²^,⁴³^,⁴⁴^,⁴⁵ The discordance between lung cavitation and sputum smear positivity observed in this study warrants consideration. Although cavitation on chest radiograph was significantly different between two patient groups, no corresponding difference in sputum smear grade was detected between groups. While cavitation is classically associated with high bacillary burden and smear positivity, this relationship is not invariable. Sputum smear microscopy has well-recognized sensitivity limitations.⁴⁶ Furthermore, smear positivity depends not only on cavity presence but also on bronchial communication, cavity drainage, and local host immune responses.¹⁸^,⁴⁷^,⁴⁸

Lineage-stratified GWAS identified three additional SNPs of interest. rs4886299 A>C in LINC00378 was significantly associated with unfavorable treatment outcome specifically among patients in the main group infected with MTB L1. Another SNP in LINC00378 has previously been associated with reduced 28-day survival in sepsis patients.⁴⁹ In patients infected with MTB L2, two additional SNPs were identified: rs74896300 G>A in EPHA3 and rs372684410 C>T in IGFLR1, both significantly associated with unfavorable treatment outcome in the main group. IGFLR1, a member of the tumor necrosis factor receptor superfamily, has been shown to correlate with CD4⁺ T cell and dendritic cell infiltration in lung adenocarcinoma and colorectal cancer.⁵⁰ EPHA3, which binds ephrin-A ligands, is expressed in mesenchymal tissues and is involved in neovascular formation in both normal adult tissue⁵¹ and solid tumors.⁵²

Previous works demonstrated that different MTB lineages maintain stable associations with human populations of distinct geographic origins. This relationship persists despite global migration and the co-circulation of diverse strains in urban settings.⁵³ The identification of host SNPs associated with TB phenotypes caused by specific MTB genotypes, but not others, is consistent with the concept of long-term co-evolution between MTB and human populations.⁵⁴ This helps explain the inability to replicate TB-associated SNPs across global populations. Similar observations have been reported in several recent studies.⁵⁵^,⁵⁶

This is the first study to comprehensively investigate TB phenotypes in a population co-infected with both MTB L1, an ancestral lineage, and the modern lineage L2. Our analysis indicates that pulmonary cavitation is predominantly determined by host genetic factors, MTB genetic clustering is primarily associated with bacterial lineage, and treatment outcomes are influenced by both host and bacterial factors as well as environmental factors such as nutritional status, co-morbidities, socioeconomic stress, overcrowding, and air pollution, which were not controlled in the present study. Several novel SNPs and candidate genes with potential relevance to TB pathogenesis were identified, though their functional roles in TB have not been characterized. These correlations were identified in the main patient group, which was genetically homogeneous and predominantly of Thai ethnicity. The confirmation of the COLEC10 association in the remaining group is encouraging; however, the lack of replication of other SNPs in this group may reflect population-specific effects, greater genetic heterogeneity, smaller sample size, or differential environmental exposures. The identified SNPs may serve as useful leads for future functional and mechanistic studies of TB pathogenesis.

Limitations of the study

This study utilized a population-based design, enrolling every isolate and patient whose samples could be collected. However, due to the relatively remote geographic location of the studied sites, only a fraction of samples was available for study, although no systematic bias was identified. The samples represented diverse MTB genotypes and multiple host genetic groups from a single province in Thailand. However, its scope remains limited in several respects. The majority of bacterial isolates belonged to sublineages of L1 and L2. The majority of L4 isolates were restricted to L4.4 and L4.5, which are considered as specialist genotype predominant in China and neighboring countries.⁵⁷ The main patient group was predominantly composed of ethnic Thais speaking Tai-Kadai languages, while the remaining group consisted largely of Tibeto-Burman speakers, primarily of Akha or Lahu ethnicity. Expansion of this work to additional populations and geographic regions is clearly warranted.

Resource availability

Lead contact

Requests for further information, resources, and reagents should be directed to and will be fulfilled by the lead contact, Prasit Palittapongarnpim (prasit.pal@mahidol.ac.th).

Materials availability

This study did not generate new unique reagents.

Data and code availability

•
The whole genome sequences of the 1,170 M. tuberculosis isolates (HAP1) and 790 isolates (HAP2) in this study are publicly available in the SRA database with accession number ERP006738 and in the BioProject database under accession code PRJNA833906, respectively.
•
This paper does not report original code.
•
Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Acknowledgments

We are grateful to the TB/HIV Research Foundation and Chiangrai Prachanukroh Hospital for collecting sample and patient data. This work was funded mainly under the Science and Technology Research Partnership for Sustainable Development (SATREPS) project (grant no. JP17jm0110010) by AMED, the Japan International Cooperation Agency (JICA), and the Department of Medical Sciences, Ministry of Public Health, Thailand and E-Asia Joint Research Program (JRA-CO-2564-13620-TH). P.T. and P.P. were supported by the Mahidol University Research Center grant and the Mahidol University Strategic Research Fund (MU-SRF-WC-02B/66).

Author contributions

S.M., K.T., P.P., and P.T. conceived the study. P.T. and P.P. analyzed, interpreted, and wrote the manuscript. S.M., P. Kunhapan, M.B., K.C., A.W., and S.W. analyzed the human genetic sequences. A.D. and W.S. performed the MTB sequencing. W.R., Y.T., and A.D. analyzed the MTB sequences. H.Y., S.N., W.I., P. Kantipong, and B.C. prepared the samples and collected the patient information. R.M., J.P., T.G.C., M.L.H., and K.T. guided the analysis. All authors approved the final version of the manuscript.

Declaration of interests

The authors declare no competing interests.

Declaration of generative AI and AI-assisted technologies in the writing process

In the preparation of this manuscript, BioRender was utilized to assist in the design and generation of graphical elements, including the graphical abstract. In addition, Claude was used exclusively for language editing purposes, including grammar checking, and improvement of sentence structure to enhance clarity, readability, and academic style. All scientific content and analyzes were conceived, verified, and approved by the authors.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Chemicals

Middlebrook 7H9 Medium	Sigma-Adlrich	Cat. #M0178
Lowenstein-Jensen medium	BD (Becton Dickinson)	Cat. # 22138
Lysozyme	Sigma-Adlrich	Cat. #L4919

Deposited data

Whole genome sequences data of the 1,170 MTB isolates of HAP1 cohort	This study	ERP006738
Whole genome sequences data of the 790 MTB isolates of HAP2 cohort	This study	PRJNA833906

Software and algorithms

Trimmomatic v0.39	Bolger et al.⁵⁸	https://github.com/usadellab/Trimmomatic/releases
BWA-MEM	Li et al.⁵⁹	https://github.com/lh3/bwa
GATK HaplotypeCaller	Poplin et al.⁶⁰	https://github.com/broadinstitute/gatk
IQ-TREE	Minh et al.⁶¹	https://www.iqtree.org
MEGA7 program	Kumar et al.⁶²	https://www.megasoftware.net/
PopArt	Leigh et al.⁶³	https://github.com/jessicawleigh/popart
PLINK version 1.90b5.4	Chang et al.⁶⁴	https://github.com/chrchang/plink-ng
Minimac4	Das et al.⁶⁵	https://github.com/statgen/Minimac4
IPCAPS version 1.1.8	Chaichoompu et al.⁶⁶	https://github.com/kridsadakorn/ipcaps
ADMIXTURE version 1.3.0	Alexander et al.⁶⁷	https://github.com/NovembreLab/admixture
FaST-LMM	Lipper et al.⁶⁸	https://github.com/fastlmm/FaST-LMM

Open in a new tab

Experimental model and study participant details

Human participants and study setting

The study was conducted in Chiang Rai Province, Northern Thailand, a region bordering Myanmar and Laos and is also a transportation hub with China via the Mekong River. This population-based study enrolled PTB patients across two distinct time periods: January 2003 to December 2010 (designated as HAP1) and January 2015 to April 2020 (HAP2). The inclusion and exclusion criteria for TB patients in the HAP1 project have been previously stated in¹⁴ and were similar to those for HAP2, with the exception that HIV-seropositive patients were excluded from HAP2. MTB was isolated from sputum samples and cultured on Lowenstein-Jensen medium or 7H9 liquid medium. All cultivable isolates were included in the study, regardless of Ziehl-Neelsen staining results. Acid-fast bacilli (AFB) in the sputum was graded as negative, 1+, 2+, or 3+.

The population of Chiang Rai was approximately 1.3 million in 2024, with a TB incidence rate of 111 per 100,000. The majority of the population consisted of native Thais, with potential genetic intermixed with Mon-Khmer and Chinese ancestries. Ethnic minorities constitute a considerable proportion of the Chiang Rai population and are collectively refer to as hill tribes. These minorities are categorized into six major groups based on spoken languages, Akha, Lahu, Karen and Lisu (all belonging to Tibeto-Burman language family), Hmong and Iu-Mien (both belonging to Hmong-Mien language family). Patient ethnicity was recorded based on self-identification.

Demographic and clinical characteristics of enrolled TB patients were presented using descriptive statistics. The demographic data included age, sex, and ethnicity. Clinical data comprised sputum AFB results, chest X-ray finding, HIV status, anti-TB drug resistance and treatment outcomes, which were defined according to WHO criteria.⁶⁹ Patients were classified into treatment failed, cured, treatment completed, died and lost to follow-up. Favorable outcome was defined as the sum of cured and treatment-completed patients. The unfavorable outcome comprised patients who died and treatment failure. Patients who were transferred out or had missing outcome data were classified as missing and excluded from multivariate and GWAS analyses. Pulmonary cavitation was assessed by chest radiography and evaluated by certified radiologists.

Ethics statement

The study was approved by the Ethical Committees of Chiang Rai Prachanukroh Hospital, Chiang Rai and the Ministry of Public Health, Thailand, and, the Institutional Review Board from Faculty of Dentistry/Faculty of Pharmacy of Mahidol University (MU-DT/PY-IRB 2022/051.2610). All participants provided written informed consent and confidentiality of their information was maintained using the participant’s code numbers. The research conformed to the principles of the Helsinki Declaration.

Method details

Bacterial genetic data

Whole genome sequencing (WGS) and phylogenetic analysis

Mycobacterial genomic DNA was extracted and prepared for WGS analysis as previously reported.⁷⁰ This study included previously reported WGS data from 1,170 MTB isolates collected during 2003–2010.¹⁴ Additional WGS of 790 MTB isolates collected during 2015–2020 was performed using Illumina NextSeq 500. Sequencing reads from both cohorts were quality-filtered using Trimmomatic v0.39 (sliding-window trimming with a window size of 4 and average read quality of at least 25).⁵⁸ Reads shorter than 50 bp were discarded. Trimmed reads were mapped to the H37Rv reference genome (GenBank accession number: NC_000962.3) using BWA-MEM v0.7.17.⁵⁹ Variant calling for each sample was performed using GATK HaplotypeCaller with a haploid model,⁶⁰ excluding bases with a quality score below 20. The following SNPs were excluded to minimize false-positive calls: (1) low-quality SNPs (filter: QD < 2 or MQ < 40, (2) SNPs located within repeat regions, mobile elements, phages, IS6110 or 13E12, PE/PPE genes and known drug resistant genes.

A phylogenetic tree of all 1,960 isolates was inferred under the maximum likelihood framework using IQ-TREE v2.0.5⁶¹ with ultrafast bootstrap support from 1,000 replications. The best-fit nucleotide substitution model was GTR+G4, as determined by ModelFinder.⁷¹ M. canettii and M. bovis were included as outgroups to root the tree, as all studied samples belong to M. tuberculosis lineages L1- L4, from which M. bovis was separated from the common ancestor more recently than M. canettii. The H37Rv reference genome (NC_000962.3) was also included.

The genotype of each bacterial isolate was assigned based on the sublineage-specific SNPs. The schemes for sublineages of M. tuberculosis lineage 1 (L1) and L2 were based on the schemes proposed by Netikul² and Thawornwatana³ respectively. Those of L3 and L4 were based on the scheme proposed by Napier.⁷²

Bacterial genetic clustering

Pairwise single-nucleotide variant (SNV) differences between isolates were calculated using MEGA7 program.⁶² A threshold of ≤12 SNVs was applied to define pairs of isolates potentially linked by recent transmission events. Accordingly, a bacterial cluster was defined as a group of MTB isolates in which each isolate differed from at least another isolate within the group by no more than 12 SNVs. Cluster sizes were categorized into two groups based on the number of MTB isolates: small clusters (2–5 isolates) and large clusters (more than 5 isolates). A minimum spanning tree was constructed using PopArt⁶³ to visualize the relationships among MTB isolates within the clusters.

Human genetic data

Genome-wide SNP genotyping

Genome-wide SNP genotyping was performed on PTB patients recruited from both HAP1 and HAP2. Details of the HAP1 SNP genotyping have been reported previously.⁶ Human DNA samples from HAP2 were genotyped using the Illumina Infinium Asian Screening Array version 1.0 (659,184 SNPs) with human genome build 38 (hg38). Genotypic datasets of HAP1 and HAP2 were merged and converted to PLINK format (PLINK version 1.90b5.4).⁶⁴

Whole genome imputation and quality control

Genotype imputation was performed using Minimac4 software⁶⁵ with 504 individuals from East Asian populations in the 1000 Genomes Project as a reference panel.⁷³ This resulted in 70,691,875 imputed SNPs. Of these imputed SNPs, we excluded 63,311,961 that were imputed with allelic R² < 0.7 (low quality filtering). The quality control of the imputed genotypes was performed using PLINK. Genetic samples; individuals with >10% missing genotype data (--mind 0.1), genetic markers with >10% missing data across all individuals (--geno 0.1) were removed. Samples with high relatedness or duplication samples were removed using the identity-by-descent estimation (IBD) value ≥0.9 (--ibd 0.9). SNPs with minor allele frequency (MAF) < 2% (--maf 0.02) and a Hardy-Weinberg equilibrium p-value < 1E-05 (--hwe 1e-05) were excluded. Finally, the number of SNPs which passed quality control was 5,075,782; these were used for the association analysis.

Population stratification

The population in Chiang Rai is genetically diverse. To conduct GWAS in a relatively homogeneous cohort and minimize confounding from differences in genetic ancestry and population structure, we performed population stratification analyses on 1,293 TB patients using Principal Component Analysis (PCA) and admixture approaches. SNPs in high linkage disequilibrium (LD) were removed using the PLINK (--indep-pairwise 50 5 0.5). The remaining 382,884 SNPs were analyzed through iterative pruning principal component analysis using IPCAPS version 1.1.8, an SNP-based population clustering tool⁶⁶ (threshold = 0.1). Nodes that contained 10 or less individuals were classified as outliers. To visualize and characterize population ancestry, admixture analysis (ADMIXTURE version 1.3.0 with default parameters)⁶⁷ was conducted to determine the optimal number of ancestors and ancestral proportions.

Genome-wide association study (GWAS)

GWAS was performed using FaST linear mixed models (FaST-LMM).⁶⁸ Prior to each association analysis, the 5,075,782 imputed SNPs were quality control again using the same criteria described above. The identity-by-descent estimation (IBD) value was set at ≥ 0.1875 (--ibd 0.1875) in order to remove the expected relatedness samples for second- and third-degree relatives. The threshold for genome-wide significance in this study was set at p < 5 × 10⁻⁸. A Manhattan plat and quantile-quantile plot were generated by R Studio program (version RStudio-2025.05.1–513). Overview of the study design and participant collection was provided in Figure S9.

N-acetyltransferase (NAT2) genotyping in TB patients

NAT2 encodes arylamine N-acetyltransferase, a liver enzyme responsible for the metabolism of various chemicals and drugs, especially the acetylation of hydrazine-based drugs such as isoniazid. The activity of NAT2 was associated with anti-TB drug-induced liver injury in Thai and Indonesian populations.⁷⁴^,⁷⁵ Haplotype-specific PCR (HS-PCR) of the human NAT2 gene was carried out according to a previous study⁷⁶ to identify SNPs of interest including rs1801279, rs1041983, rs1801280, rs1799929, rs1799930, rs1208 and rs1799931 in the NAT2 region on chromosome 8. These are the signature SNPs for the following seven clusters: NAT2∗14, NAT2∗13, NAT2∗5, NAT2∗11, NAT2∗6, NAT2∗12, and NAT2∗7, respectively. The phenotype of NAT2 acetylator was classified as rapid acetylator (RA), intermediate acetylator (IA), or slow acetylator (SA), as described in.⁷⁶

Quantification and statistical analysis

Demographic and clinical data were recorded as continuous or categorical variables. The categorical variables were presented as frequencies and percentages. Statistical analysis was performed using the PASW Statistics 21.0 (SPSS Inc., IL, USA). The variables of TB patients in HAP1 and HAP2 groups were compared by the chi-square test or Fisher’s exact test. Variables such as age, human PCA groups, sputum AFB results, HIV status, TB drug resistance, MTB lineages and genetic clustering were included in the univariate analysis. Variables with a p-value <0.2 in the univariate analysis were included in the multivariate logistic regression analysis to identify independent predictors of treatment outcomes, lung cavitation and MTB clustering. A p-value less than 0.05 was considered statistically significant.

Footnotes

Supplemental information can be found online at https://doi.org/10.1016/j.isci.2026.117699.

Supplemental information

Document S1. Figures S1–S9, Tables S4–S6, and Data S1

mmc1.pdf^{(1.1MB, pdf)}

Table S1. Demographic and clinical profiles of TB patients in two different time periods: HAP1 group (2003–2010) and HAP2 group (2015–2020)

mmc2.xlsx^{(14.7KB, xlsx)}

Table S2. Demographic and clinical profiles of 4 major lineages and sublineages of L1, L2, and L4

mmc3.xlsx^{(22.1KB, xlsx)}

Table S3. Demographic and clinical profiles of TB patients infected with various MTB lineages during HAP1 and HAP2 periods

mmc4.xlsx^{(16.4KB, xlsx)}

Table S7. Human genetic loci associated with unfavorable outcomes, lung cavitation, and genetic clustering of MTB in the main and remaining groups of TB patients

mmc5.xlsx^{(11.9KB, xlsx)}

References

1.World Health Organization . Global Tuberculosis Report 2025. WHO; 2025. [Google Scholar]
2.Netikul T., Thawornwattana Y., Mahasirimongkol S., Yanai H., Maung H.M.W., Chongsuvivatwong V., Palittapongarnpim P. Whole-genome single nucleotide variant phylogenetic analysis of Mycobacterium tuberculosis Lineage 1 in endemic regions of Asia and Africa. Sci. Rep. 2022;12:1565. doi: 10.1038/s41598-022-05524-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Thawornwattana Y., Mahasirimongkol S., Yanai H., Maung H.M.W., Cui Z., Chongsuvivatwong V., Palittapongarnpim P. Revised nomenclature and SNP barcode for Mycobacterium tuberculosis lineage 2. Microb. Genom. 2021;7 doi: 10.1099/mgen.0.000697. [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Zhang L., Pang Y., Yu X., Wang Y., Lu J., Gao M., Huang H., Zhao Y. Risk factors for pulmonary cavitation in tuberculosis patients from China. Emerg. Microb. Infect. 2016;5 doi: 10.1038/emi.2016.111. [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Möller M., Kinnear C.J. Human global and population-specific genetic susceptibility to Mycobacterium tuberculosis infection and disease. Curr. Opin. Pulm. Med. 2020;26:302–310. doi: 10.1097/MCP.0000000000000672. [DOI] [PubMed] [Google Scholar]
6.Ramirez-Alejo N., Santos-Argumedo L. Innate defects of the IL-12/IFN-gamma axis in susceptibility to infections by mycobacteria and salmonella. J. Interferon Cytokine Res. 2014;34:307–317. doi: 10.1089/jir.2013.0050. [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Arias A.A., Neehus A.L., Ogishi M., Meynier V., Krebs A., Lazarov T., Lee A.M., Arango-Franco C.A., Yang R., Orrego J., et al. Tuberculosis in otherwise healthy adults with inherited TNF deficiency. Nature. 2024;633:417–425. doi: 10.1038/s41586-024-07866-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Boisson-Dupuis S., Ramirez-Alejo N., Li Z., Patin E., Rao G., Kerner G., Lim C.K., Krementsov D.N., Hernandez N., Ma C.S., et al. Tuberculosis and impaired IL-23-dependent IFN-gamma immunity in humans homozygous for a common TYK2 missense variant. Sci. Immunol. 2018;3 doi: 10.1126/sciimmunol.aau8714. [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Ogishi M., Yang R., Rodriguez R., Golec D.P., Martin E., Philippot Q., Bohlen J., Pelham S.J., Arias A.A., Khan T., et al. Inherited human ITK deficiency impairs IFN-gamma immunity and underlies tuberculosis. J. Exp. Med. 2023;220 doi: 10.1084/jem.20220484. [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Dashian M.A., Shipulin G.A., Deviatkin A.A. Genetic Susceptibility to Tuberculosis and the Utility of Polygenic Scores in Population Stratification. Int. J. Mol. Sci. 2025;26 doi: 10.3390/ijms26199544. [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Palittapongarnpim P., Tantivitayakul P., Aiewsakun P., Mahasirimongkol S., Jaemsai B. Genomic Interactions Between Mycobacterium tuberculosis and Humans. Annu. Rev. Genom. Hum. Genet. 2024;25:183–209. doi: 10.1146/annurev-genom-021623-101844. [DOI] [PubMed] [Google Scholar]
12.Phelan J., Gomez-Gonzalez P.J., Andreu N., Omae Y., Toyo-Oka L., Yanai H., Miyahara R., Nedsuwan S., de Sessions P.F., Campino S., et al. Genome-wide host-pathogen analyses reveal genetic interaction points in tuberculosis disease. Nat. Commun. 2023;14:549. doi: 10.1038/s41467-023-36282-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Omae Y., Toyo-Oka L., Yanai H., Nedsuwan S., Wattanapokayakit S., Satproedprai N., Smittipat N., Palittapongarnpim P., Sawanpanyalert P., Inunchot W., et al. Pathogen lineage-based genome-wide association study identified CD53 as susceptible locus in tuberculosis. J. Hum. Genet. 2017;62:1015–1022. doi: 10.1038/jhg.2017.82. [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Ajawatanawong P., Yanai H., Smittipat N., Disratthakit A., Yamada N., Miyahara R., Nedsuwan S., Imasanguan W., Kantipong P., Chaiyasirinroje B., et al. A novel Ancestral Beijing sublineage of Mycobacterium tuberculosis suggests the transition site to Modern Beijing sublineages. Sci. Rep. 2019;9 doi: 10.1038/s41598-019-50078-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Teeter L.D., Ha N.P., Ma X., Wenger J., Cronin W.A., Musser J.M., Graviss E.A. Evaluation of large genotypic Mycobacterium tuberculosis clusters: contributions from remote and recent transmission. Tuberculosis. 2013;93:S38–S46. doi: 10.1016/S1472-9792(13)70009-X. [DOI] [PubMed] [Google Scholar]
16.Glynn J.R., Crampin A.C., Traore H., Chaguluka S., Mwafulirwa D.T., Alghamdi S., Ngwira B.M.M., Yates M.D., Drobniewski F.D., Fine P.E.M. Determinants of cluster size in large, population-based molecular epidemiology study of tuberculosis, northern Malawi. Emerg. Infect. Dis. 2008;14:1060–1066. doi: 10.3201/eid1407.060468. [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Turner R.D., Chiu C., Churchyard G.J., Esmail H., Lewinsohn D.M., Gandhi N.R., Fennelly K.P. Tuberculosis Infectiousness and Host Susceptibility. J. Infect. Dis. 2017;216:S636–S643. doi: 10.1093/infdis/jix361. [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Urbanowski M.E., Ordonez A.A., Ruiz-Bedoya C.A., Jain S.K., Bishai W.R. Cavitary tuberculosis: the gateway of disease transmission. Lancet Infect. Dis. 2020;20:e117–e128. doi: 10.1016/S1473-3099(20)30148-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Holt K.E., McAdam P., Thai P.V.K., Thuong N.T.T., Ha D.T.M., Lan N.N., Lan N.H., Nhu N.T.Q., Hai H.T., Ha V.T.N., et al. Frequent transmission of the Mycobacterium tuberculosis Beijing lineage and positive selection for the EsxW Beijing variant in Vietnam. Nat. Genet. 2018;50:849–856. doi: 10.1038/s41588-018-0117-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Liu Q., Ma A., Wei L., Pang Y., Wu B., Luo T., Zhou Y., Zheng H.X., Jiang Q., Gan M., et al. China's tuberculosis epidemic stems from historical expansion of four strains of Mycobacterium tuberculosis. Nat. Ecol. Evol. 2018;2:1982–1992. doi: 10.1038/s41559-018-0680-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Freschi L., Vargas R., Jr., Husain A., Kamal S.M.M., Skrahina A., Tahseen S., Ismail N., Barbova A., Niemann S., Cirillo D.M., et al. Population structure, biogeography and transmissibility of Mycobacterium tuberculosis. Nat. Commun. 2021;12:6099. doi: 10.1038/s41467-021-26248-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Troegeler A., Lugo-Villarino G., Hansen S., Rasolofo V., Henriksen M.L., Mori K., Ohtani K., Duval C., Mercier I., Bénard A., et al. Collectin CL-LK Is a Novel Soluble Pattern Recognition Receptor for Mycobacterium tuberculosis. PLoS One. 2015;10 doi: 10.1371/journal.pone.0132692. [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Menzies N.A., Allwood B.W., Dean A.S., Dodd P.J., Houben R.M.G.J., James L.P., Knight G.M., Meghji J., Nguyen L.N., Rachow A., et al. Global burden of disease due to rifampicin-resistant tuberculosis: a mathematical modeling analysis. Nat. Commun. 2023;14:6182. doi: 10.1038/s41467-023-41937-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Seung K.J., Gelmanova I.E., Peremitin G.G., Golubchikova V.T., Pavlova V.E., Sirotkina O.B., Yanova G.V., Strelis A.K. The effect of initial drug resistance on treatment response and acquired drug resistance during standardized short-course chemotherapy for tuberculosis. Clin. Infect. Dis. 2004;39:1321–1328. doi: 10.1086/425005. [DOI] [PubMed] [Google Scholar]
25.Villegas L., Otero L., Sterling T.R., Huaman M.A., Van der Stuyft P., Gotuzzo E., Seas C. Prevalence, Risk Factors, and Treatment Outcomes of Isoniazid- and Rifampicin-Mono-Resistant Pulmonary Tuberculosis in Lima, Peru. PLoS One. 2016;11 doi: 10.1371/journal.pone.0152933. [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Azuma J., Ohno M., Kubota R., Yokota S., Nagai T., Tsuyuguchi K., Okuda Y., Takashima T., Kamimura S., Fujio Y., et al. NAT2 genotype guided regimen reduces isoniazid-induced liver injury and early treatment failure in the 6-month four-drug standard treatment of tuberculosis: a randomized controlled trial for pharmacogenetics-based therapy. Eur. J. Clin. Pharmacol. 2013;69:1091–1101. doi: 10.1007/s00228-012-1429-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Langmuir AD W.S., Pope A.S., POPE A.S. Certain Factors Influencing the Mortality of Sanatorium Treated Cases of Pulmonary Tuberculosis. N. Engl. J. Med. 1934;211:26–29. doi: 10.1056/nejm193407052110106. [DOI] [Google Scholar]
28.Smith J.D., Meehan M.H., Crean J., McCann A. Alpha T-catenin (CTNNA3): a gene in the hand is worth two in the nest. Cell. Mol. Life Sci. 2011;68:2493–2498. doi: 10.1007/s00018-011-0728-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Chiarella S.E., Rabin E.E., Ostilla L.A., Flozak A.S., Gottardi C.J. alphaT-catenin: A developmentally dispensable, disease-linked member of the alpha-catenin family. Tissue Barriers. 2018;6 doi: 10.1080/21688370.2018.1463896. [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Mokrousov I., Pasechnik O., Vyazovaya A., Yarusova I., Gerasimova A., Blokh A., Zhuravlev V. Impact of pathobiological diversity of Mycobacterium tuberculosis on clinical features and lethal outcome of tuberculosis. BMC Microbiol. 2022;22:50. doi: 10.1186/s12866-022-02461-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Hang N.T.L., Maeda S., Keicho N., Thuong P.H., Endo H. Sublineages of Mycobacterium tuberculosis Beijing genotype strains and unfavorable outcomes of anti-tuberculosis treatment. Tuberculosis. 2015;95:336–342. doi: 10.1016/j.tube.2015.02.040. [DOI] [PubMed] [Google Scholar]
32.Klopper M., Heupink T.H., Hill-Cawthorne G., Streicher E.M., Dippenaar A., de Vos M., Abdallah A.M., Limberis J., Merker M., Burns S., et al. A landscape of genomic alterations at the root of a near-untreatable tuberculosis epidemic. BMC Med. 2020;18:24. doi: 10.1186/s12916-019-1487-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Wilcke J.T., Askgaard D.S., Nybo Jensen B., Døssing M. Radiographic spectrum of adult pulmonary tuberculosis in a developed country. Respir. Med. 1998;92:493–497. doi: 10.1016/s0954-6111(98)90297-9. [DOI] [PubMed] [Google Scholar]
34.Wampande E.M., Mupere E., Debanne S.M., Asiimwe B.B., Nsereko M., Mayanja H., Eisenach K., Kaplan G., Boom H.W., Gagneux S., Joloba M.L. Long-term dominance of Mycobacterium tuberculosis Uganda family in peri-urban Kampala-Uganda is not associated with cavitary disease. BMC Infect. Dis. 2013;13:484. doi: 10.1186/1471-2334-13-484. [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Sun J., Siroy A., Lokareddy R.K., Speer A., Doornbos K.S., Cingolani G., Niederweis M. The tuberculosis necrotizing toxin kills macrophages by hydrolyzing NAD. Nat. Struct. Mol. Biol. 2015;22:672–678. doi: 10.1038/nsmb.3064. [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Verma S., Bhatt K., Lovey A., Ribeiro-Rodrigues R., Durbin J., Jones-López E.C., Palaci M., Vinhas S.A., Alland D., Dietze R., et al. Transmission phenotype of Mycobacterium tuberculosis strains is mechanistically linked to induction of distinct pulmonary pathology. PLoS Pathog. 2019;15 doi: 10.1371/journal.ppat.1007613. [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Coussens A.K., Wilkinson R.J., Nikolayevskyy V., Elkington P.T., Hanifa Y., Islam K., Timms P.M., Bothamley G.H., Claxton A.P., Packe G.E., et al. Ethnic variation in inflammatory profile in tuberculosis. PLoS Pathog. 2013;9 doi: 10.1371/journal.ppat.1003468. [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Vijayaraghavan M., Gadad S.S., Dhandayuthapani S. Long non-coding RNA transcripts in Mycobacterium tuberculosis-host interactions. Noncoding. RNA Res. 2025;11:281–293. doi: 10.1016/j.ncrna.2024.12.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Zhang X., Chen C., Xu Y. Long Non-coding RNAs in Tuberculosis: From Immunity to Biomarkers. Front. Microbiol. 2022;13 doi: 10.3389/fmicb.2022.883513. [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Grandjean L., Monteserin J., Gilman R., Pauschardt J., Rokadiya S., Bonilla C., Ritacco V., Vidal J.R., Parkhill J., Peacock S., et al. Association between bacterial homoplastic variants and radiological pathology in tuberculosis. Thorax. 2020;75:584–591. doi: 10.1136/thoraxjnl-2019-213281. [DOI] [PMC free article] [PubMed] [Google Scholar]
41.Almeida F.M., Ventura T.L.B., Amaral E.P., Ribeiro S.C.M., Calixto S.D., Manhães M.R., Rezende A.L., Souza G.S., de Carvalho I.S., Silva E.C., et al. Hypervirulent Mycobacterium tuberculosis strain triggers necrotic lung pathology associated with enhanced recruitment of neutrophils in resistant C57BL/6 mice. PLoS One. 2017;12 doi: 10.1371/journal.pone.0173715. [DOI] [PMC free article] [PubMed] [Google Scholar]
42.Kato-Maeda M., Shanley C.A., Ackart D., Jarlsberg L.G., Shang S., Obregon-Henao A., Harton M., Basaraba R.J., Henao-Tamayo M., Barrozo J.C., et al. Beijing sublineages of Mycobacterium tuberculosis differ in pathogenicity in the guinea pig. Clin. Vaccine Immunol. 2012;19:1227–1237. doi: 10.1128/CVI.00250-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Ates L.S., Dippenaar A., Ummels R., Piersma S.R., van der Woude A.D., van der Kuij K., Le Chevalier F., Mata-Espinosa D., Barrios-Payán J., Marquina-Castillo B., et al. Mutations in ppe38 block PE_PGRS secretion and increase virulence of Mycobacterium tuberculosis. Nat. Microbiol. 2018;3:181–188. doi: 10.1038/s41564-017-0090-6. [DOI] [PubMed] [Google Scholar]
44.Koleske B., Schill C., Rajagopalan S., Shee S., Martinez-Martinez Y.B., Gupta M., Shen J., Jacobs W.R., Jr., Bishai W.R. Loss of the PPE71-esxX-esxY-PPE38 locus drives adaptive transcriptional responses and hypervirulence of Mycobacterium tuberculosis lineage 2. Sci. Adv. 2025;11 doi: 10.1126/sciadv.adw5194. [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Bottai D., Frigui W., Sayes F., Di Luca M., Spadoni D., Pawlik A., Zoppo M., Orgeur M., Khanna V., Hardy D., et al. TbD1 deletion as a driver of the evolutionary success of modern epidemic Mycobacterium tuberculosis lineages. Nat. Commun. 2020;11:684. doi: 10.1038/s41467-020-14508-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Steingart K.R., Ng V., Henry M., Hopewell P.C., Ramsay A., Cunningham J., Urbanczik R., Perkins M.D., Aziz M.A., Pai M. Sputum processing methods to improve the sensitivity of smear microscopy for tuberculosis: a systematic review. Lancet Infect. Dis. 2006;6:664–674. doi: 10.1016/S1473-3099(06)70602-8. [DOI] [PubMed] [Google Scholar]
47.Ozkaya S., Bilgin S., Findik S., Kök H.C., Yuksel C., Atıcı A.G. Endobronchial tuberculosis: histopathological subsets and microbiological results. Multidiscip. Respir. Med. 2012;7:34. doi: 10.1186/2049-6958-7-34. [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Nolan A., Fajardo E., Huie M.L., Condos R., Pooran A., Dawson R., Dheda K., Bateman E., Rom W.N., Weiden M.D. Increased production of IL-4 and IL-12p40 from bronchoalveolar lavage cells are biomarkers of Mycobacterium tuberculosis in the sputum. PLoS One. 2013;8 doi: 10.1371/journal.pone.0059461. [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Hernandez-Beeftink T., Guillen-Guio B., Lorenzo-Salazar J.M., Corrales A., Suarez-Pajes E., Feng R., Rubio-Rodríguez L.A., Paynton M.L., Cruz R., García-Laorden M.I., et al. A genome-wide association study of survival in patients with sepsis. Crit. Care. 2022;26:341. doi: 10.1186/s13054-022-04208-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
50.Jin R., Du F., Han X., Guo J., Song W., Xia Y., Yue X., Yang D., Tong J., Zhang Q., Liu Y. Prognostic Value of Insulin Growth Factor-Like Receptor 1 (IGFLR1) in Stage II and III Colorectal Cancer and Its Association with Immune Cell Infiltration. Appl. Biochem. Biotechnol. 2025;197:427–442. doi: 10.1007/s12010-024-05006-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
51.To C., Farnsworth R.H., Vail M.E., Chheang C., Gargett C.E., Murone C., Llerena C., Major A.T., Scott A.M., Janes P.W., Lackmann M. Hypoxia-controlled EphA3 marks a human endometrium-derived multipotent mesenchymal stromal cell that supports vascular growth. PLoS One. 2014;9 doi: 10.1371/journal.pone.0112106. [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Vail M.E., Murone C., Tan A., Hii L., Abebe D., Janes P.W., Lee F.T., Baer M., Palath V., Bebbington C., et al. Targeting EphA3 inhibits cancer growth by disrupting the tumor stromal microenvironment. Cancer Res. 2014;74:4470–4481. doi: 10.1158/0008-5472.CAN-14-0218. [DOI] [PubMed] [Google Scholar]
53.Hirsh A.E., Tsolaki A.G., DeRiemer K., Feldman M.W., Small P.M. Stable association between strains of Mycobacterium tuberculosis and their human host populations. Proc. Natl. Acad. Sci. USA. 2004;101:4871–4876. doi: 10.1073/pnas.0305627101. [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Brites D., Gagneux S. Co-evolution of Mycobacterium tuberculosis and Homo sapiens. Immunol. Rev. 2015;264:6–24. doi: 10.1111/imr.12264. [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Schurz H., Naranbhai V., Yates T.A., Gilchrist J.J., Parks T., Dodd P.J., Möller M., Hoal E.G., Morris A.P., Hill A.V.S., International Tuberculosis Host Genetics Consortium Multi-ancestry meta-analysis of host genetic susceptibility to tuberculosis identifies shared genetic architecture. eLife. 2024;13 doi: 10.7554/eLife.84394. [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Jiao L., Song J., Chen H., Gu D., Liao S., Zhou J., Wu T., Zhang J., Hu X., Bai H., et al. Genetic architecture of tuberculosis susceptibility: A comprehensive research synopsis, meta-analyses, and epidemiological evidence. Infect. Genet. Evol. 2022;104 doi: 10.1016/j.meegid.2022.105352. [DOI] [PubMed] [Google Scholar]
57.Stucki D., Brites D., Jeljeli L., Coscolla M., Liu Q., Trauner A., Fenner L., Rutaihwa L., Borrell S., Luo T., et al. Mycobacterium tuberculosis lineage 4 comprises globally distributed and geographically restricted sublineages. Nat. Genet. 2016;48:1535–1543. doi: 10.1038/ng.3704. [DOI] [PMC free article] [PubMed] [Google Scholar]
58.Bolger A.M., Lohse M., Usadel B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics. 2014;30:2114–2120. doi: 10.1093/bioinformatics/btu170. [DOI] [PMC free article] [PubMed] [Google Scholar]
59.Li H., Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics. 2009;25:1754–1760. doi: 10.1093/bioinformatics/btp324. [DOI] [PMC free article] [PubMed] [Google Scholar]
60.Poplin R., Chang P.C., Alexander D., Schwartz S., Colthurst T., Ku A., Newburger D., Dijamco J., Nguyen N., Afshar P.T., et al. A universal SNP and small-indel variant caller using deep neural networks. Nat. Biotechnol. 2018;36:983–987. doi: 10.1038/nbt.4235. [DOI] [PubMed] [Google Scholar]
61.Minh B.Q., Schmidt H.A., Chernomor O., Schrempf D., Woodhams M.D., von Haeseler A., Lanfear R. IQ-TREE 2: New Models and Efficient Methods for Phylogenetic Inference in the Genomic Era. Mol. Biol. Evol. 2020;37:1530–1534. doi: 10.1093/molbev/msaa015. [DOI] [PMC free article] [PubMed] [Google Scholar]
62.Kumar S., Stecher G., Tamura K. MEGA7: Molecular Evolutionary Genetics Analysis Version 7.0 for Bigger Datasets. Mol. Biol. Evol. 2016;33:1870–1874. doi: 10.1093/molbev/msw054. [DOI] [PMC free article] [PubMed] [Google Scholar]
63.Leigh J.W., Bryant D. Popart: full-feature software for haplotype network construction. Methods Ecol. Evol. 2015;6:1110–1116. doi: 10.1111/2041-210x.12410. [DOI] [Google Scholar]
64.Chang C.C., Chow C.C., Tellier L.C., Vattikuti S., Purcell S.M., Lee J.J. Second-generation PLINK: rising to the challenge of larger and richer datasets. GigaScience. 2015;4:7. doi: 10.1186/s13742-015-0047-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
65.Das S., Forer L., Schönherr S., Sidore C., Locke A.E., Kwong A., Vrieze S.I., Chew E.Y., Levy S., McGue M., et al. Next-generation genotype imputation service and methods. Nat. Genet. 2016;48:1284–1287. doi: 10.1038/ng.3656. [DOI] [PMC free article] [PubMed] [Google Scholar]
66.Chaichoompu K., Abegaz F., Tongsima S., Shaw P.J., Sakuntabhai A., Pereira L., Van Steen K. IPCAPS: an R package for iterative pruning to capture population structure. Source Code Biol. Med. 2019;14:2. doi: 10.1186/s13029-019-0072-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
67.Alexander D.H., Novembre J., Lange K. Fast model-based estimation of ancestry in unrelated individuals. Genome Res. 2009;19:1655–1664. doi: 10.1101/gr.094052.109. [DOI] [PMC free article] [PubMed] [Google Scholar]
68.Lippert C., Listgarten J., Liu Y., Kadie C.M., Davidson R.I., Heckerman D. FaST linear mixed models for genome-wide association studies. Nat. Methods. 2011;8:833–835. doi: 10.1038/nmeth.1681. [DOI] [PubMed] [Google Scholar]
69.World Health Organization . WHO; 2013. Definitions and Reporting Framework for Tuberculosis. [Google Scholar]
70.Palittapongarnpim P., Ajawatanawong P., Viratyosin W., Smittipat N., Disratthakit A., Mahasirimongkol S., Yanai H., Yamada N., Nedsuwan S., Imasanguan W., et al. Evidence for Host-Bacterial Co-evolution via Genome Sequence Analysis of 480 Thai Mycobacterium tuberculosis Lineage 1 Isolates. Sci. Rep. 2018;8 doi: 10.1038/s41598-018-29986-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
71.Kalyaanamoorthy S., Minh B.Q., Wong T.K.F., von Haeseler A., Jermiin L.S. ModelFinder: fast model selection for accurate phylogenetic estimates. Nat. Methods. 2017;14:587–589. doi: 10.1038/nmeth.4285. [DOI] [PMC free article] [PubMed] [Google Scholar]
72.Napier G., Campino S., Merid Y., Abebe M., Woldeamanuel Y., Aseffa A., Hibberd M.L., Phelan J., Clark T.G. Robust barcoding and identification of Mycobacterium tuberculosis lineages for epidemiological and clinical studies. Genome Med. 2020;12:114. doi: 10.1186/s13073-020-00817-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
73.Auton A., Abecasis G.R., Altshuler D.M., Co-Chair, Durbin R.M., Abecasis G.R., Bentley D.R., Chakravarti A., Clark A.G., Donnelly P., Eichler E.E., et al. A global reference for human genetic variation. Nature. 2015;526:68–74. doi: 10.1038/nature15393. [DOI] [PMC free article] [PubMed] [Google Scholar]
74.Suvichapanich S., Wattanapokayakit S., Mushiroda T., Yanai H., Chuchottawon C., Kantima T., Nedsuwan S., Suwankesawong W., Sonsupap C., Pannarunothai R., et al. Genomewide Association Study Confirming the Association of NAT2 with Susceptibility to Antituberculosis Drug-Induced Liver Injury in Thai Patients. Antimicrob. Agents Chemother. 2019;63 doi: 10.1128/AAC.02692-18. [DOI] [PMC free article] [PubMed] [Google Scholar]
75.Yuliwulandari R., Susilowati R.W., Wicaksono B.D., Viyati K., Prayuni K., Razari I., Kristin E., Syafrizal S., Sri Diana E., Sri Diana E., et al. NAT2 variants are associated with drug-induced liver injury caused by anti-tuberculosis drugs in Indonesian patients with tuberculosis. J. Hum. Genet. 2016;61:533–537. doi: 10.1038/jhg.2016.10. [DOI] [PubMed] [Google Scholar]
76.Wichukchinda N., Pakdee J., Kunhapan P., Imunchot W., Toyo-Oka L., Tokunaga K., Mahasirimongkol S. Haplotype-specific PCR for NAT2 diplotyping. Hum. Genome Var. 2020;7:13. doi: 10.1038/s41439-020-0101-7. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Document S1. Figures S1–S9, Tables S4–S6, and Data S1

mmc1.pdf^{(1.1MB, pdf)}

Table S1. Demographic and clinical profiles of TB patients in two different time periods: HAP1 group (2003–2010) and HAP2 group (2015–2020)

mmc2.xlsx^{(14.7KB, xlsx)}

Table S2. Demographic and clinical profiles of 4 major lineages and sublineages of L1, L2, and L4

mmc3.xlsx^{(22.1KB, xlsx)}

Table S3. Demographic and clinical profiles of TB patients infected with various MTB lineages during HAP1 and HAP2 periods

mmc4.xlsx^{(16.4KB, xlsx)}

Table S7. Human genetic loci associated with unfavorable outcomes, lung cavitation, and genetic clustering of MTB in the main and remaining groups of TB patients

mmc5.xlsx^{(11.9KB, xlsx)}

Data Availability Statement

•
The whole genome sequences of the 1,170 M. tuberculosis isolates (HAP1) and 790 isolates (HAP2) in this study are publicly available in the SRA database with accession number ERP006738 and in the BioProject database under accession code PRJNA833906, respectively.
•
This paper does not report original code.
•
Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

[bib1] 1.World Health Organization . Global Tuberculosis Report 2025. WHO; 2025. [Google Scholar]

[bib2] 2.Netikul T., Thawornwattana Y., Mahasirimongkol S., Yanai H., Maung H.M.W., Chongsuvivatwong V., Palittapongarnpim P. Whole-genome single nucleotide variant phylogenetic analysis of Mycobacterium tuberculosis Lineage 1 in endemic regions of Asia and Africa. Sci. Rep. 2022;12:1565. doi: 10.1038/s41598-022-05524-0. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib3] 3.Thawornwattana Y., Mahasirimongkol S., Yanai H., Maung H.M.W., Cui Z., Chongsuvivatwong V., Palittapongarnpim P. Revised nomenclature and SNP barcode for Mycobacterium tuberculosis lineage 2. Microb. Genom. 2021;7 doi: 10.1099/mgen.0.000697. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib4] 4.Zhang L., Pang Y., Yu X., Wang Y., Lu J., Gao M., Huang H., Zhao Y. Risk factors for pulmonary cavitation in tuberculosis patients from China. Emerg. Microb. Infect. 2016;5 doi: 10.1038/emi.2016.111. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib5] 5.Möller M., Kinnear C.J. Human global and population-specific genetic susceptibility to Mycobacterium tuberculosis infection and disease. Curr. Opin. Pulm. Med. 2020;26:302–310. doi: 10.1097/MCP.0000000000000672. [DOI] [PubMed] [Google Scholar]

[bib6] 6.Ramirez-Alejo N., Santos-Argumedo L. Innate defects of the IL-12/IFN-gamma axis in susceptibility to infections by mycobacteria and salmonella. J. Interferon Cytokine Res. 2014;34:307–317. doi: 10.1089/jir.2013.0050. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib7] 7.Arias A.A., Neehus A.L., Ogishi M., Meynier V., Krebs A., Lazarov T., Lee A.M., Arango-Franco C.A., Yang R., Orrego J., et al. Tuberculosis in otherwise healthy adults with inherited TNF deficiency. Nature. 2024;633:417–425. doi: 10.1038/s41586-024-07866-3. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib8] 8.Boisson-Dupuis S., Ramirez-Alejo N., Li Z., Patin E., Rao G., Kerner G., Lim C.K., Krementsov D.N., Hernandez N., Ma C.S., et al. Tuberculosis and impaired IL-23-dependent IFN-gamma immunity in humans homozygous for a common TYK2 missense variant. Sci. Immunol. 2018;3 doi: 10.1126/sciimmunol.aau8714. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib9] 9.Ogishi M., Yang R., Rodriguez R., Golec D.P., Martin E., Philippot Q., Bohlen J., Pelham S.J., Arias A.A., Khan T., et al. Inherited human ITK deficiency impairs IFN-gamma immunity and underlies tuberculosis. J. Exp. Med. 2023;220 doi: 10.1084/jem.20220484. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib10] 10.Dashian M.A., Shipulin G.A., Deviatkin A.A. Genetic Susceptibility to Tuberculosis and the Utility of Polygenic Scores in Population Stratification. Int. J. Mol. Sci. 2025;26 doi: 10.3390/ijms26199544. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib11] 11.Palittapongarnpim P., Tantivitayakul P., Aiewsakun P., Mahasirimongkol S., Jaemsai B. Genomic Interactions Between Mycobacterium tuberculosis and Humans. Annu. Rev. Genom. Hum. Genet. 2024;25:183–209. doi: 10.1146/annurev-genom-021623-101844. [DOI] [PubMed] [Google Scholar]

[bib12] 12.Phelan J., Gomez-Gonzalez P.J., Andreu N., Omae Y., Toyo-Oka L., Yanai H., Miyahara R., Nedsuwan S., de Sessions P.F., Campino S., et al. Genome-wide host-pathogen analyses reveal genetic interaction points in tuberculosis disease. Nat. Commun. 2023;14:549. doi: 10.1038/s41467-023-36282-w. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib13] 13.Omae Y., Toyo-Oka L., Yanai H., Nedsuwan S., Wattanapokayakit S., Satproedprai N., Smittipat N., Palittapongarnpim P., Sawanpanyalert P., Inunchot W., et al. Pathogen lineage-based genome-wide association study identified CD53 as susceptible locus in tuberculosis. J. Hum. Genet. 2017;62:1015–1022. doi: 10.1038/jhg.2017.82. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib14] 14.Ajawatanawong P., Yanai H., Smittipat N., Disratthakit A., Yamada N., Miyahara R., Nedsuwan S., Imasanguan W., Kantipong P., Chaiyasirinroje B., et al. A novel Ancestral Beijing sublineage of Mycobacterium tuberculosis suggests the transition site to Modern Beijing sublineages. Sci. Rep. 2019;9 doi: 10.1038/s41598-019-50078-3. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib15] 15.Teeter L.D., Ha N.P., Ma X., Wenger J., Cronin W.A., Musser J.M., Graviss E.A. Evaluation of large genotypic Mycobacterium tuberculosis clusters: contributions from remote and recent transmission. Tuberculosis. 2013;93:S38–S46. doi: 10.1016/S1472-9792(13)70009-X. [DOI] [PubMed] [Google Scholar]

[bib16] 16.Glynn J.R., Crampin A.C., Traore H., Chaguluka S., Mwafulirwa D.T., Alghamdi S., Ngwira B.M.M., Yates M.D., Drobniewski F.D., Fine P.E.M. Determinants of cluster size in large, population-based molecular epidemiology study of tuberculosis, northern Malawi. Emerg. Infect. Dis. 2008;14:1060–1066. doi: 10.3201/eid1407.060468. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib17] 17.Turner R.D., Chiu C., Churchyard G.J., Esmail H., Lewinsohn D.M., Gandhi N.R., Fennelly K.P. Tuberculosis Infectiousness and Host Susceptibility. J. Infect. Dis. 2017;216:S636–S643. doi: 10.1093/infdis/jix361. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib18] 18.Urbanowski M.E., Ordonez A.A., Ruiz-Bedoya C.A., Jain S.K., Bishai W.R. Cavitary tuberculosis: the gateway of disease transmission. Lancet Infect. Dis. 2020;20:e117–e128. doi: 10.1016/S1473-3099(20)30148-1. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib19] 19.Holt K.E., McAdam P., Thai P.V.K., Thuong N.T.T., Ha D.T.M., Lan N.N., Lan N.H., Nhu N.T.Q., Hai H.T., Ha V.T.N., et al. Frequent transmission of the Mycobacterium tuberculosis Beijing lineage and positive selection for the EsxW Beijing variant in Vietnam. Nat. Genet. 2018;50:849–856. doi: 10.1038/s41588-018-0117-9. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib20] 20.Liu Q., Ma A., Wei L., Pang Y., Wu B., Luo T., Zhou Y., Zheng H.X., Jiang Q., Gan M., et al. China's tuberculosis epidemic stems from historical expansion of four strains of Mycobacterium tuberculosis. Nat. Ecol. Evol. 2018;2:1982–1992. doi: 10.1038/s41559-018-0680-6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib21] 21.Freschi L., Vargas R., Jr., Husain A., Kamal S.M.M., Skrahina A., Tahseen S., Ismail N., Barbova A., Niemann S., Cirillo D.M., et al. Population structure, biogeography and transmissibility of Mycobacterium tuberculosis. Nat. Commun. 2021;12:6099. doi: 10.1038/s41467-021-26248-1. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib22] 22.Troegeler A., Lugo-Villarino G., Hansen S., Rasolofo V., Henriksen M.L., Mori K., Ohtani K., Duval C., Mercier I., Bénard A., et al. Collectin CL-LK Is a Novel Soluble Pattern Recognition Receptor for Mycobacterium tuberculosis. PLoS One. 2015;10 doi: 10.1371/journal.pone.0132692. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib23] 23.Menzies N.A., Allwood B.W., Dean A.S., Dodd P.J., Houben R.M.G.J., James L.P., Knight G.M., Meghji J., Nguyen L.N., Rachow A., et al. Global burden of disease due to rifampicin-resistant tuberculosis: a mathematical modeling analysis. Nat. Commun. 2023;14:6182. doi: 10.1038/s41467-023-41937-9. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib24] 24.Seung K.J., Gelmanova I.E., Peremitin G.G., Golubchikova V.T., Pavlova V.E., Sirotkina O.B., Yanova G.V., Strelis A.K. The effect of initial drug resistance on treatment response and acquired drug resistance during standardized short-course chemotherapy for tuberculosis. Clin. Infect. Dis. 2004;39:1321–1328. doi: 10.1086/425005. [DOI] [PubMed] [Google Scholar]

[bib25] 25.Villegas L., Otero L., Sterling T.R., Huaman M.A., Van der Stuyft P., Gotuzzo E., Seas C. Prevalence, Risk Factors, and Treatment Outcomes of Isoniazid- and Rifampicin-Mono-Resistant Pulmonary Tuberculosis in Lima, Peru. PLoS One. 2016;11 doi: 10.1371/journal.pone.0152933. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib26] 26.Azuma J., Ohno M., Kubota R., Yokota S., Nagai T., Tsuyuguchi K., Okuda Y., Takashima T., Kamimura S., Fujio Y., et al. NAT2 genotype guided regimen reduces isoniazid-induced liver injury and early treatment failure in the 6-month four-drug standard treatment of tuberculosis: a randomized controlled trial for pharmacogenetics-based therapy. Eur. J. Clin. Pharmacol. 2013;69:1091–1101. doi: 10.1007/s00228-012-1429-9. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib27] 27.Langmuir AD W.S., Pope A.S., POPE A.S. Certain Factors Influencing the Mortality of Sanatorium Treated Cases of Pulmonary Tuberculosis. N. Engl. J. Med. 1934;211:26–29. doi: 10.1056/nejm193407052110106. [DOI] [Google Scholar]

[bib28] 28.Smith J.D., Meehan M.H., Crean J., McCann A. Alpha T-catenin (CTNNA3): a gene in the hand is worth two in the nest. Cell. Mol. Life Sci. 2011;68:2493–2498. doi: 10.1007/s00018-011-0728-0. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib29] 29.Chiarella S.E., Rabin E.E., Ostilla L.A., Flozak A.S., Gottardi C.J. alphaT-catenin: A developmentally dispensable, disease-linked member of the alpha-catenin family. Tissue Barriers. 2018;6 doi: 10.1080/21688370.2018.1463896. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib30] 30.Mokrousov I., Pasechnik O., Vyazovaya A., Yarusova I., Gerasimova A., Blokh A., Zhuravlev V. Impact of pathobiological diversity of Mycobacterium tuberculosis on clinical features and lethal outcome of tuberculosis. BMC Microbiol. 2022;22:50. doi: 10.1186/s12866-022-02461-w. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib31] 31.Hang N.T.L., Maeda S., Keicho N., Thuong P.H., Endo H. Sublineages of Mycobacterium tuberculosis Beijing genotype strains and unfavorable outcomes of anti-tuberculosis treatment. Tuberculosis. 2015;95:336–342. doi: 10.1016/j.tube.2015.02.040. [DOI] [PubMed] [Google Scholar]

[bib32] 32.Klopper M., Heupink T.H., Hill-Cawthorne G., Streicher E.M., Dippenaar A., de Vos M., Abdallah A.M., Limberis J., Merker M., Burns S., et al. A landscape of genomic alterations at the root of a near-untreatable tuberculosis epidemic. BMC Med. 2020;18:24. doi: 10.1186/s12916-019-1487-2. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib33] 33.Wilcke J.T., Askgaard D.S., Nybo Jensen B., Døssing M. Radiographic spectrum of adult pulmonary tuberculosis in a developed country. Respir. Med. 1998;92:493–497. doi: 10.1016/s0954-6111(98)90297-9. [DOI] [PubMed] [Google Scholar]

[bib34] 34.Wampande E.M., Mupere E., Debanne S.M., Asiimwe B.B., Nsereko M., Mayanja H., Eisenach K., Kaplan G., Boom H.W., Gagneux S., Joloba M.L. Long-term dominance of Mycobacterium tuberculosis Uganda family in peri-urban Kampala-Uganda is not associated with cavitary disease. BMC Infect. Dis. 2013;13:484. doi: 10.1186/1471-2334-13-484. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib35] 35.Sun J., Siroy A., Lokareddy R.K., Speer A., Doornbos K.S., Cingolani G., Niederweis M. The tuberculosis necrotizing toxin kills macrophages by hydrolyzing NAD. Nat. Struct. Mol. Biol. 2015;22:672–678. doi: 10.1038/nsmb.3064. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib36] 36.Verma S., Bhatt K., Lovey A., Ribeiro-Rodrigues R., Durbin J., Jones-López E.C., Palaci M., Vinhas S.A., Alland D., Dietze R., et al. Transmission phenotype of Mycobacterium tuberculosis strains is mechanistically linked to induction of distinct pulmonary pathology. PLoS Pathog. 2019;15 doi: 10.1371/journal.ppat.1007613. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib37] 37.Coussens A.K., Wilkinson R.J., Nikolayevskyy V., Elkington P.T., Hanifa Y., Islam K., Timms P.M., Bothamley G.H., Claxton A.P., Packe G.E., et al. Ethnic variation in inflammatory profile in tuberculosis. PLoS Pathog. 2013;9 doi: 10.1371/journal.ppat.1003468. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib38] 38.Vijayaraghavan M., Gadad S.S., Dhandayuthapani S. Long non-coding RNA transcripts in Mycobacterium tuberculosis-host interactions. Noncoding. RNA Res. 2025;11:281–293. doi: 10.1016/j.ncrna.2024.12.005. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib39] 39.Zhang X., Chen C., Xu Y. Long Non-coding RNAs in Tuberculosis: From Immunity to Biomarkers. Front. Microbiol. 2022;13 doi: 10.3389/fmicb.2022.883513. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib40] 40.Grandjean L., Monteserin J., Gilman R., Pauschardt J., Rokadiya S., Bonilla C., Ritacco V., Vidal J.R., Parkhill J., Peacock S., et al. Association between bacterial homoplastic variants and radiological pathology in tuberculosis. Thorax. 2020;75:584–591. doi: 10.1136/thoraxjnl-2019-213281. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib41] 41.Almeida F.M., Ventura T.L.B., Amaral E.P., Ribeiro S.C.M., Calixto S.D., Manhães M.R., Rezende A.L., Souza G.S., de Carvalho I.S., Silva E.C., et al. Hypervirulent Mycobacterium tuberculosis strain triggers necrotic lung pathology associated with enhanced recruitment of neutrophils in resistant C57BL/6 mice. PLoS One. 2017;12 doi: 10.1371/journal.pone.0173715. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib42] 42.Kato-Maeda M., Shanley C.A., Ackart D., Jarlsberg L.G., Shang S., Obregon-Henao A., Harton M., Basaraba R.J., Henao-Tamayo M., Barrozo J.C., et al. Beijing sublineages of Mycobacterium tuberculosis differ in pathogenicity in the guinea pig. Clin. Vaccine Immunol. 2012;19:1227–1237. doi: 10.1128/CVI.00250-12. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib43] 43.Ates L.S., Dippenaar A., Ummels R., Piersma S.R., van der Woude A.D., van der Kuij K., Le Chevalier F., Mata-Espinosa D., Barrios-Payán J., Marquina-Castillo B., et al. Mutations in ppe38 block PE_PGRS secretion and increase virulence of Mycobacterium tuberculosis. Nat. Microbiol. 2018;3:181–188. doi: 10.1038/s41564-017-0090-6. [DOI] [PubMed] [Google Scholar]

[bib44] 44.Koleske B., Schill C., Rajagopalan S., Shee S., Martinez-Martinez Y.B., Gupta M., Shen J., Jacobs W.R., Jr., Bishai W.R. Loss of the PPE71-esxX-esxY-PPE38 locus drives adaptive transcriptional responses and hypervirulence of Mycobacterium tuberculosis lineage 2. Sci. Adv. 2025;11 doi: 10.1126/sciadv.adw5194. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib45] 45.Bottai D., Frigui W., Sayes F., Di Luca M., Spadoni D., Pawlik A., Zoppo M., Orgeur M., Khanna V., Hardy D., et al. TbD1 deletion as a driver of the evolutionary success of modern epidemic Mycobacterium tuberculosis lineages. Nat. Commun. 2020;11:684. doi: 10.1038/s41467-020-14508-5. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib46] 46.Steingart K.R., Ng V., Henry M., Hopewell P.C., Ramsay A., Cunningham J., Urbanczik R., Perkins M.D., Aziz M.A., Pai M. Sputum processing methods to improve the sensitivity of smear microscopy for tuberculosis: a systematic review. Lancet Infect. Dis. 2006;6:664–674. doi: 10.1016/S1473-3099(06)70602-8. [DOI] [PubMed] [Google Scholar]

[bib47] 47.Ozkaya S., Bilgin S., Findik S., Kök H.C., Yuksel C., Atıcı A.G. Endobronchial tuberculosis: histopathological subsets and microbiological results. Multidiscip. Respir. Med. 2012;7:34. doi: 10.1186/2049-6958-7-34. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib48] 48.Nolan A., Fajardo E., Huie M.L., Condos R., Pooran A., Dawson R., Dheda K., Bateman E., Rom W.N., Weiden M.D. Increased production of IL-4 and IL-12p40 from bronchoalveolar lavage cells are biomarkers of Mycobacterium tuberculosis in the sputum. PLoS One. 2013;8 doi: 10.1371/journal.pone.0059461. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib49] 49.Hernandez-Beeftink T., Guillen-Guio B., Lorenzo-Salazar J.M., Corrales A., Suarez-Pajes E., Feng R., Rubio-Rodríguez L.A., Paynton M.L., Cruz R., García-Laorden M.I., et al. A genome-wide association study of survival in patients with sepsis. Crit. Care. 2022;26:341. doi: 10.1186/s13054-022-04208-5. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib50] 50.Jin R., Du F., Han X., Guo J., Song W., Xia Y., Yue X., Yang D., Tong J., Zhang Q., Liu Y. Prognostic Value of Insulin Growth Factor-Like Receptor 1 (IGFLR1) in Stage II and III Colorectal Cancer and Its Association with Immune Cell Infiltration. Appl. Biochem. Biotechnol. 2025;197:427–442. doi: 10.1007/s12010-024-05006-1. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib51] 51.To C., Farnsworth R.H., Vail M.E., Chheang C., Gargett C.E., Murone C., Llerena C., Major A.T., Scott A.M., Janes P.W., Lackmann M. Hypoxia-controlled EphA3 marks a human endometrium-derived multipotent mesenchymal stromal cell that supports vascular growth. PLoS One. 2014;9 doi: 10.1371/journal.pone.0112106. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib52] 52.Vail M.E., Murone C., Tan A., Hii L., Abebe D., Janes P.W., Lee F.T., Baer M., Palath V., Bebbington C., et al. Targeting EphA3 inhibits cancer growth by disrupting the tumor stromal microenvironment. Cancer Res. 2014;74:4470–4481. doi: 10.1158/0008-5472.CAN-14-0218. [DOI] [PubMed] [Google Scholar]

[bib53] 53.Hirsh A.E., Tsolaki A.G., DeRiemer K., Feldman M.W., Small P.M. Stable association between strains of Mycobacterium tuberculosis and their human host populations. Proc. Natl. Acad. Sci. USA. 2004;101:4871–4876. doi: 10.1073/pnas.0305627101. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib54] 54.Brites D., Gagneux S. Co-evolution of Mycobacterium tuberculosis and Homo sapiens. Immunol. Rev. 2015;264:6–24. doi: 10.1111/imr.12264. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib55] 55.Schurz H., Naranbhai V., Yates T.A., Gilchrist J.J., Parks T., Dodd P.J., Möller M., Hoal E.G., Morris A.P., Hill A.V.S., International Tuberculosis Host Genetics Consortium Multi-ancestry meta-analysis of host genetic susceptibility to tuberculosis identifies shared genetic architecture. eLife. 2024;13 doi: 10.7554/eLife.84394. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib56] 56.Jiao L., Song J., Chen H., Gu D., Liao S., Zhou J., Wu T., Zhang J., Hu X., Bai H., et al. Genetic architecture of tuberculosis susceptibility: A comprehensive research synopsis, meta-analyses, and epidemiological evidence. Infect. Genet. Evol. 2022;104 doi: 10.1016/j.meegid.2022.105352. [DOI] [PubMed] [Google Scholar]

[bib57] 57.Stucki D., Brites D., Jeljeli L., Coscolla M., Liu Q., Trauner A., Fenner L., Rutaihwa L., Borrell S., Luo T., et al. Mycobacterium tuberculosis lineage 4 comprises globally distributed and geographically restricted sublineages. Nat. Genet. 2016;48:1535–1543. doi: 10.1038/ng.3704. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib58] 58.Bolger A.M., Lohse M., Usadel B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics. 2014;30:2114–2120. doi: 10.1093/bioinformatics/btu170. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib59] 59.Li H., Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics. 2009;25:1754–1760. doi: 10.1093/bioinformatics/btp324. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib60] 60.Poplin R., Chang P.C., Alexander D., Schwartz S., Colthurst T., Ku A., Newburger D., Dijamco J., Nguyen N., Afshar P.T., et al. A universal SNP and small-indel variant caller using deep neural networks. Nat. Biotechnol. 2018;36:983–987. doi: 10.1038/nbt.4235. [DOI] [PubMed] [Google Scholar]

[bib61] 61.Minh B.Q., Schmidt H.A., Chernomor O., Schrempf D., Woodhams M.D., von Haeseler A., Lanfear R. IQ-TREE 2: New Models and Efficient Methods for Phylogenetic Inference in the Genomic Era. Mol. Biol. Evol. 2020;37:1530–1534. doi: 10.1093/molbev/msaa015. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib62] 62.Kumar S., Stecher G., Tamura K. MEGA7: Molecular Evolutionary Genetics Analysis Version 7.0 for Bigger Datasets. Mol. Biol. Evol. 2016;33:1870–1874. doi: 10.1093/molbev/msw054. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib63] 63.Leigh J.W., Bryant D. Popart: full-feature software for haplotype network construction. Methods Ecol. Evol. 2015;6:1110–1116. doi: 10.1111/2041-210x.12410. [DOI] [Google Scholar]

[bib64] 64.Chang C.C., Chow C.C., Tellier L.C., Vattikuti S., Purcell S.M., Lee J.J. Second-generation PLINK: rising to the challenge of larger and richer datasets. GigaScience. 2015;4:7. doi: 10.1186/s13742-015-0047-8. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib65] 65.Das S., Forer L., Schönherr S., Sidore C., Locke A.E., Kwong A., Vrieze S.I., Chew E.Y., Levy S., McGue M., et al. Next-generation genotype imputation service and methods. Nat. Genet. 2016;48:1284–1287. doi: 10.1038/ng.3656. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib66] 66.Chaichoompu K., Abegaz F., Tongsima S., Shaw P.J., Sakuntabhai A., Pereira L., Van Steen K. IPCAPS: an R package for iterative pruning to capture population structure. Source Code Biol. Med. 2019;14:2. doi: 10.1186/s13029-019-0072-6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib67] 67.Alexander D.H., Novembre J., Lange K. Fast model-based estimation of ancestry in unrelated individuals. Genome Res. 2009;19:1655–1664. doi: 10.1101/gr.094052.109. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib68] 68.Lippert C., Listgarten J., Liu Y., Kadie C.M., Davidson R.I., Heckerman D. FaST linear mixed models for genome-wide association studies. Nat. Methods. 2011;8:833–835. doi: 10.1038/nmeth.1681. [DOI] [PubMed] [Google Scholar]

[bib69] 69.World Health Organization . WHO; 2013. Definitions and Reporting Framework for Tuberculosis. [Google Scholar]

[bib70] 70.Palittapongarnpim P., Ajawatanawong P., Viratyosin W., Smittipat N., Disratthakit A., Mahasirimongkol S., Yanai H., Yamada N., Nedsuwan S., Imasanguan W., et al. Evidence for Host-Bacterial Co-evolution via Genome Sequence Analysis of 480 Thai Mycobacterium tuberculosis Lineage 1 Isolates. Sci. Rep. 2018;8 doi: 10.1038/s41598-018-29986-3. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib71] 71.Kalyaanamoorthy S., Minh B.Q., Wong T.K.F., von Haeseler A., Jermiin L.S. ModelFinder: fast model selection for accurate phylogenetic estimates. Nat. Methods. 2017;14:587–589. doi: 10.1038/nmeth.4285. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib72] 72.Napier G., Campino S., Merid Y., Abebe M., Woldeamanuel Y., Aseffa A., Hibberd M.L., Phelan J., Clark T.G. Robust barcoding and identification of Mycobacterium tuberculosis lineages for epidemiological and clinical studies. Genome Med. 2020;12:114. doi: 10.1186/s13073-020-00817-3. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib73] 73.Auton A., Abecasis G.R., Altshuler D.M., Co-Chair, Durbin R.M., Abecasis G.R., Bentley D.R., Chakravarti A., Clark A.G., Donnelly P., Eichler E.E., et al. A global reference for human genetic variation. Nature. 2015;526:68–74. doi: 10.1038/nature15393. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib74] 74.Suvichapanich S., Wattanapokayakit S., Mushiroda T., Yanai H., Chuchottawon C., Kantima T., Nedsuwan S., Suwankesawong W., Sonsupap C., Pannarunothai R., et al. Genomewide Association Study Confirming the Association of NAT2 with Susceptibility to Antituberculosis Drug-Induced Liver Injury in Thai Patients. Antimicrob. Agents Chemother. 2019;63 doi: 10.1128/AAC.02692-18. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib75] 75.Yuliwulandari R., Susilowati R.W., Wicaksono B.D., Viyati K., Prayuni K., Razari I., Kristin E., Syafrizal S., Sri Diana E., Sri Diana E., et al. NAT2 variants are associated with drug-induced liver injury caused by anti-tuberculosis drugs in Indonesian patients with tuberculosis. J. Hum. Genet. 2016;61:533–537. doi: 10.1038/jhg.2016.10. [DOI] [PubMed] [Google Scholar]

[bib76] 76.Wichukchinda N., Pakdee J., Kunhapan P., Imunchot W., Toyo-Oka L., Tokunaga K., Mahasirimongkol S. Haplotype-specific PCR for NAT2 diplotyping. Hum. Genome Var. 2020;7:13. doi: 10.1038/s41439-020-0101-7. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Association between human genomes and mycobacterial genomes contributes to severity and transmissibility of pulmonary tuberculosis

Pornpen Tantivitayakul

Punna Kunhapan

Manon Boonbangyang

Kridsadakorn Chaichoompu

Alisa Wilantho

Waritta Sawaengdee

Wuthiwat Ruangchai

Yuttapong Thawornwattana

Hideki Yanai

Sukanya Wattanapokayakit

Areeya Disratthakit

Reiko Miyahara

Supalert Nedsuwan

Worarat Imasanguan

Pacharee Kantipong

Boonchai Chaiyasirinroje

Jody Phelan

Taane G Clark

Martin L Hibberd

Katsushi Tokunaga

Surakameth Mahasirimongkol

Prasit Palittapongarnpim

Summary

Graphical abstract

Highlights

Introduction

Results

Demographic and clinical characteristics of TB patients

MTB genotypic diversity and transmission clustering

Table 1.

Human host population structure

Figure 1.

Table 2.

Risk factors for MTB clustering, treatment outcome, and pulmonary cavitation

MTB genetic clustering

Table 3.

Treatment outcome

Table 4.

Pulmonary cavitation

Table 5.

GWAS of host genetic loci in relation to MTB phenotypes

Genome-wide significant SNPs across the main group

Figure 2.

Table 6.

Lineage-stratified GWAS

Figure 3.

Candidate gene association in the remaining group

Discussion

Limitations of the study

Resource availability

Lead contact

Materials availability

Data and code availability

Acknowledgments

Author contributions

Declaration of interests

Declaration of generative AI and AI-assisted technologies in the writing process

STAR★Methods

Key resources table

Experimental model and study participant details

Human participants and study setting

Ethics statement

Method details

Bacterial genetic data

Whole genome sequencing (WGS) and phylogenetic analysis

Bacterial genetic clustering

Human genetic data

Genome-wide SNP genotyping

Whole genome imputation and quality control

Population stratification

Genome-wide association study (GWAS)

N-acetyltransferase (NAT2) genotyping in TB patients

Quantification and statistical analysis

Footnotes

Supplemental information

References

Associated Data

Supplementary Materials