Whole genome sequencing from clinical primary Mycobacterium tuberculosis liquid cultures: pushing the boundaries
A Dippenaar(1) E Costa Conceição(2) F Wells(2) A Paulse(2) T H Heupink(1) V Rennie(1) M De Diego Fuertes(1) R M Warren(2) A Van Rie(1)
1:University of Antwerp; 2:Stellenbosch University
Whole genome sequencing (WGS) for personalized treatment should be accurate and have a short turnaround time. Until WGS directly from sputum samples is feasible, clinical primary liquid cultures (CPLCs) are an attractive alternative. Mycobacterium tuberculosis CPLCs often produce low-quality DNA due to contaminants and low biomass. We submitted 158 CTAB-extracted DNA samples from CPLCs for WGS at a commercial service provider. Of the 158 samples, 52 (33%) passed quality control (QC) with a DNA Integrity Number (DIN) ≥7, 83 (53%) had a DIN of 6-6.9 (‘on hold’) and 23 (15%) a DIN <6 (‘failed)’. Because DNA replacement would require a subculture and create a delay of 3-8 weeks, we reviewed the fragment size distribution, peak height, and graph shape of electropherograms of 106 samples with DINs <7. We decided to process 104 of 106 (98%) samples with DIN<7. Of the 156 samples processed by library prep, 137 passed library QC, 16 had low library concentrations, and three completely failed library QC. WGS was performed on all except the three samples that completely failed. Using the MAGMA pipeline only three of 153 (2%) analyses failed (coverage depth <15×). The other 150 (98%) samples had a median depth of coverage of 1002× and a median of 91% mapped reads. MAGMA generated an actionable drug-resistance profile for all 150 samples that passed WGS-QC. Our results demonstrate that the standard criteria for DNA and library QC may be too stringent as actionable genomic drug resistance profiles from almost all (95%) CPLC samples were obtained.