Point prevalence surveys of acute infection presentation and antibiotic prescribing in selected primary healthcare facilities in North-West and Gauteng provinces of South Africa

ples.Mixed infection (minority strain 3% in Deeplex or any level in MAGMA) occurred in 3 samples.One sample was reported as 91% Mtb 2.2.1 + 8% Mtb 4.4.1.1 by MAGMA, and 87% M. bovis + 39% undefined Mtb lineage by Deeplex.One sample was reported as 99% NTM + 1% Mtb by MAGMA but 100% Mtb 4.3 by Deeplex.Another sample contained Mtb lineages 1, 2, 7, 4, 5, animal strains and M. canettii (4%) + undefined Mtb lineage (96%) in Deeplex and 100% Mtb 4.1.1.3 in MAGMA.The fourth sample contained two undefined lineages at 11% and 89% in Deeplex but 100% Mtb 2.2.1 in MAGMA.According to Deeplex, 33/53 samples shared an identical spoligotype (7 groups of size 2-19) and 20/53 had a unique spoligotype.Based on a 12 SNP cut-off, 4 samples were part of a WGS cluster and 55/59 were unclustered.Of the spoligotype groups, only one group of the 7 groups at least 2 identical spoligotypes had a SNP distance < 12. Strain typing for strains identified by MAGMA as lineage 4 were varied in the Deeplex analysis, apart from strains identified by both as sublineage 4.3Discussion: A significant proportion of the samples analysed by Deeplex were reported as "Other than lineage 4.9".This definition is not informative for strain typing.The resolution of Deeplex is less granular than WGS; it failed to assign sublineage for approximately two-thirds of the samples for which it could assign lineage.Deeplex particularly struggles with strain typing for lineage 4 strains unless they are lineage 4.3 strains. Conclusion:Our findings highlight poor phylogenetic resolution of tNGS, limiting its ability to inform public health interventions.