Strengths and limitations

This study is the first to perform head-to-head comparison between MIRU-VNTR and WGS quantitatively using hospital-based data in Taiwan. Most of the studies perform the analysis of VNTR first, then they use WGS to analyze the result of MIRU-VNTR [10, 26]. In these studies, they reported that WGS had higher resolution by using some validation data, and it was helpful to exclude transmission events identified by MIRU-VNTR. Nevertheless, it is seldom to analyze all samples by two genotyped methods simultaneously in one study probably due to limited resource and cost. As a result, the result of clusters always lacks the part which is identified by WGS, but MIRU considers not as the cluster. Besides, the addresses of activity places in our study were not only residential, but also more places were recorded, such as work. The more detailed information of activity places is recorded, the more we are possible to understand the potential transmission location.

There were still a few potential limitations of this study. Our population was collected from the specific hospital system, not all TB patients of Kaohsiung were included in 2017,

which may cause selection bias. Besides, the study period was only nearly one year, it was difficult to understand complete transmission in Kaohsiung. These problems might lead to an underestimation of clusters of TB transmission. However, we planned to conduct a population-based study to collect all cases from all hospitals in Kaohsiung for over one year in the future. Our epidemiological data based on contact investigation was difficult to record all the epidemiological links. Hence, epidemiological links were possibly missed some parts of links between cases, and it also had the problem of recall bias when interviewing patients. Several limitations of WGS technologies and analysis tools were challenged that the workflows had various criteria, and lacked standardization and validation. There was no international standard for the cut-off of SNP, although many studies use the number of SNPs to define WGS cluster. The SNPs cut-off as adequate cluster definition still needed to take some factors into account, such as the TB incidence, the characteristic and mutation rates of MTB in this area, the duration of transmission, and so on.

In conclusion, with decreasing costs of next-generation sequencing (NGS) technologies, WGS will become an essential tool in routine use of identifying clusters.

Moreover, our results also support that the cases of WGS clusters match more with epidemiological information than clusters by MIRU-VNTR. Although WGS has potential

ability to replace routine typing MIRU-VNTR [13], some challenges remain such as how to compare among different studies when WGS applying, and how to better define clusters of recent or reactive TB transmission by SNP thresholds in different conditionally area. Besides, workflow of WGS has various criteria, and lacks standardization and validation. It should be validated and established an algorithm as general standard for meeting public health needs in the future.

References

1. Thain, N., et al., Towards better prediction of Mycobacterium tuberculosis lineages from MIRU-VNTR data. Infection, Genetics and Evolution, 2018.

2. Supply, P., et al., Proposal for Standardization of Optimized Mycobacterial Interspersed Repetitive Unit-Variable-Number Tandem Repeat Typing of Mycobacterium tuberculosis. 2006. 44(12): p. 4498-4510.

3. Gurjav, U., et al., Temporal dynamics of Mycobacterium tuberculosis genotypes in New South Wales, Australia. BMC Infect Dis, 2014. 14: p. 455.

4. Zamani, S., et al., MIRU-VNTR analysis of the Mycobacterium tuberculosis isolates from three provinces of Iran. Scand J Infect Dis, 2013. 45(2): p. 124-30.

5. Shamputa, I.C., et al., Genetic diversity of Mycobacterium tuberculosis isolates from a tertiary care tuberculosis hospital in South Korea. Journal of clinical microbiology, 2010. 48(2): p. 387-394.

6. Perez Del Molino Bernal, I.C., et al., Genomic Diversity of Mycobacterium tuberculosis Complex Strains in Cantabria (Spain), a Moderate TB Incidence Setting. PLoS One, 2016. 11(6): p. e0157266.

7. Wyllie, D.H., et al., A Quantitative Evaluation of MIRU-VNTR Typing Against Whole-Genome Sequencing for Identifying Mycobacterium tuberculosis

Transmission: A Prospective Observational Cohort Study. EBioMedicine, 2018.

34: p. 122-130.

8. Comas, I., et al., Genotyping of Genetically Monomorphic Bacteria: DNA Sequencing in Mycobacterium tuberculosis Highlights the Limitations of Current Methodologies. PLOS ONE, 2009. 4(11): p. e7815.

9. Luo, T., et al., Combination of Single Nucleotide Polymorphism and Variable-Number Tandem Repeats for Genotyping a Homogenous Population of Mycobacterium tuberculosis Beijing Strains in China. Journal of Clinical Microbiology, 2012. 50(3): p. 633.

10. Stucki, D., et al., Standard Genotyping Overestimates Transmission of

Mycobacterium tuberculosis among Immigrants in a Low-Incidence Country. J Clin Microbiol, 2016. 54(7): p. 1862-1870.

11. Jamieson, F.B., et al., Whole-genome sequencing of the Mycobacterium tuberculosis Manila sublineage results in less clustering and better resolution than mycobacterial interspersed repetitive-unit-variable-number tandem-repeat (MIRU-VNTR) typing and spoligotyping. J Clin Microbiol, 2014. 52(10): p.

3795-8.

12. Gardy, J.L., et al., Whole-Genome Sequencing and Social-Network Analysis of a Tuberculosis Outbreak. 2011. 364(8): p. 730-739.

13. Walker, T.M., et al., Whole-genome sequencing to delineate Mycobacterium tuberculosis outbreaks: a retrospective observational study. The Lancet.

Infectious diseases, 2013. 13(2): p. 137-146.

14. Jajou, R., et al., Epidemiological links between tuberculosis cases identified twice as efficiently by whole genome sequencing than conventional molecular typing: A population-based study. PLoS One, 2018. 13(4): p. e0195413.

15. Koster, K.J., et al., Genomic sequencing is required for identification of tuberculosis transmission in Hawaii. BMC Infect Dis, 2018. 18(1): p. 608.

16. Stimson, J., et al., Beyond the SNP Threshold: Identifying Outbreak Clusters Using Inferred Transmissions. Molecular Biology and Evolution, 2019. 36(3): p.

587-603.

17. Alaridah, N., et al., Transmission dynamics study of tuberculosis isolates with whole genome sequencing in southern Sweden. Scientific Reports, 2019. 9(1): p.

4931.

18. Luo, T., et al., Development of a Hierarchical Variable-Number Tandem Repeat Typing Scheme for Mycobacterium tuberculosis in China. PLOS ONE, 2014.

9(2): p. e89726.

19. 莊佩君, 周如文, 103 年度全程研究報告-開發基因分型方法及建置結核菌資

料庫. 2014, 衛生福利部疾病管制署: 衛生福利部疾病管制署.

20. Hunter, P.R. and M.A. Gaston, Numerical index of the discriminatory ability of typing systems: an application of Simpson's index of diversity. Journal of clinical microbiology, 1988. 26(11): p. 2465-2466.

21. Sola, C., et al., Genotyping of the Mycobacterium tuberculosis complex using MIRUs: association with VNTR and spoligotyping for molecular epidemiology and evolutionary genetics. Infection, Genetics and Evolution, 2003. 3(2): p. 125-133.

22. Cole, S.T., et al., Deciphering the biology of Mycobacterium tuberculosis from the complete genome sequence. Nature, 1998. 393(6685): p. 537-544.

23. Meehan, C.J., et al., The relationship between transmission time and clustering methods in Mycobacterium tuberculosis epidemiology. 2018: p. 302232.

24. Kohl, T.A., et al., Whole-genome-based Mycobacterium tuberculosis surveillance: a standardized, portable, and expandable approach. J Clin Microbiol, 2014. 52(7): p. 2479-86.

25. Pankhurst, L.J., et al., Rapid, comprehensive, and affordable mycobacterial diagnosis with whole-genome sequencing: a prospective study. Lancet Respir Med, 2016. 4(1): p. 49-58.

26. Tyler, A.D., et al., Application of whole genome sequence analysis to the study of Mycobacterium tuberculosis in Nunavut, Canada. PLOS ONE, 2017. 12(10):

p. e0185656.

27. Sloot, R., et al., Clustering of tuberculosis cases based on variable-number tandem-repeat typing in relation to the population structure of Mycobacterium tuberculosis in the Netherlands. Journal of clinical microbiology, 2013. 51(7): p.

2427-2431.

28. Coscolla, M. and S. Gagneux, Consequences of genomic diversity in

Mycobacterium tuberculosis. Seminars in Immunology, 2014. 26(6): p. 431-444.

29. Chaisson, M., P. Pevzner, and H. Tang, Fragment assembly with short reads.

Bioinformatics, 2004. 20(13): p. 2067-2074.

30. Meehan, C.J., et al., Whole genome sequencing of Mycobacterium tuberculosis:

current standards and open issues. Nature Reviews Microbiology, 2019.

31. Pfeiffer, F., et al., Systematic evaluation of error rates and causes in short samples in next-generation sequencing. Scientific Reports, 2018. 8(1): p. 10950.

Appendix

Figure 1. Flow chart of sample collection. The samples were the isolates of TB patients.

Table 1. The baseline descriptive characteristic of patients in Kaohsiung Medical University Hospital from March 2017 to December 2017.

Age in years, n (%)

Table 2. The consistence of clustered pairs between 24-loci MIRU-VNTR and WGS from 155 isolates of patients.

The 163 patients excluding Case 18, 19, 32, 67, 77, 103, 138, 143). (+) was mean the number of pairs identified as the cluster by the tool; (−) was mean the number of pairs identified as non-cluster by the tool.

WGS (+) WGS (−) Total

8 80 88

15 11832 11847

23 11912 11935

24-loci MIRU-VNTR (+) 24-loci MIRU-VNTR (−)

Total

Figure 2. The distribution of patient locations by lineages.

Figure 3. The distribution of Cluster A patient locations.

Figure 4. The distribution of Cluster B-D patient locations.

Figure 5. Network of WGS-clustered cases.

Figure 6. Neighbor joining tree of Cluster A.

Figure 7. Network of MIRU-clustered cases with geographical correlation.

Figure 8. Neighbor joining tree of Cluster P.

Figure 9. The relationship of variation between 24-loci MIRU-VNTR and WGS.

Figure 10. The relationship of variation between 24-loci MIRU-VNTR and WGS (under 50 SNPs).

Figure 11. Venn diagram of 24-loci MIRU-VNTR and WGS typing of 155 MTB isolates from Kaohsiung and geographical correlation.

Supplement

Table S1. The characteristic of patients between WGS-clustered and Unique.

Male

Table S2. The characteristic of patients between MIRU-clustered and Unique.

Table S3. The consistence of clustered pairs between 8-loci MIRU-VNTR and WGS from 158 isolates of patients.

(+) was mean the number of pairs identified as the cluster by the tool; (−) was mean the number of pairs identified as non-cluster by the tool. Patients were excluding Case 19, 32, 77, 138, 143.

WGS (+) WGS (−) Total

8 157 165

15 12223 12238

23 12380 12403

Total

8-loci MIRU-VNTR (+) 8-loci MIRU-VNTR (−)

Table S4. The consistence of clustered pairs between 10-loci MIRU-VNTR and WGS from 158 isolates of patients.

(+) was mean the number of pairs identified as the cluster by the tool; (−) was mean the number of pairs identified as non-cluster by the tool. Patients were excluding Case 19, 32, 77, 138, 143.

WGS (+) WGS (−) Total

4 39 43

19 12341 12360

23 12380 12403

Total

10-loci MIRU-VNTR (+) 10-loci MIRU-VNTR (−)

Table S5. The consistence of clustered pairs between 28-loci MIRU-VNTR and WGS from 155 isolates of patients.

(+) was mean the number of pairs identified as the cluster by the tool; (−) was mean the number of pairs identified as non-cluster by the tool. The patients were excluding Case 18, 19, 32, 67, 77, 103, 138, 143.

WGS (+) WGS (−) Total

4 23 27

19 12044 12063

23 12067 12090

Total

28-loci MIRU-VNTR (+) 28-loci MIRU-VNTR (−)

Table S6. The consistence of clustered case number between 24-loci MIRU-VNTR and WGS from 158 isolates of patients.

(+) was mean the number of pairs identified as the cluster by the tool; (−) was mean the number of pairs identified as non-cluster by the tool. Patients were excluding Case 19, 32, 77, 138, 143.

WGS (+) WGS (−) Total

8 31 39

3 113 116

23 144 155

24-loci MIRU-VNTR (+) 24-loci MIRU-VNTR (−)

Total

Table S7. The Hunter Gaston discriminatory index of each loci.

Bold was the locus of 24-loci international standard.

Loci Loci Loci

Table S8. The number of pairs in different SNP-definition.

(+) was mean the number of pairs identified as the cluster by the tool; (−) was mean the number of pairs identified as non-cluster by the tool.

82 81 80

Figure S1. The relationship of variation between 8-loci MIRU-VNTR and WGS.

Figure S2. The relationship of variation between 10-loci MIRU-VNTR and WGS.

Figure S3. The relationship of variation between 28-loci MIRU-VNTR and WGS.

Figure S4. The relationship of variation between MIRU-VNTR and WGS in Lineage 1 (N=56).

Figure S5. The relationship of variation between MIRU-VNTR and WGS in Lineage 2 (N=63).

Figure S6. The relationship of variation between MIRU-VNTR and WGS in Lineage 4 (N=36).

Figure S7. Network of MIRU-clustered cases.

Figure S8. The number of pairs in the different group of SNPs.