Genomic profiling in acute myeloid leukemia with complex karyotype

Anna opened her talk by emphasising how improving outcomes in acute myleoid leukaemia (AML) remains a major challenge. AML is "a very heterogeneous disease", characterised by a variety of complex karyotypic changes, with patients being stratified for treatment according to the genetic abnormalities present, such as translocations. Traditional methods of karyotyping are slow and laborious, typically taking one to two weeks; this clearly prevents rapid treatment decisions, and so "performing karyotyping in a short time frame would be desirable".

By performing shallow whole-genome sequencing using the Oxford Nanopore GridION, Anna and her team were able to obtain karyotype information for 20 AML patient samples within only 24 hours. This workflow involved: library preparation (Ligation Sequencing Kit SQK-LSK109, 60 min) of unfragmented DNA, GridION sequencing (1 flow cell per patient, overnight 18 hour run), and bioinformatics analysis (1 hour). The throughput of one flow cell was typically 10-14 Gb, with a peak read-length of 28 kbp. Two analysis pipelines were applied, both firstly using ngmlr for sequence alignment to hg19 and samtools for indexing; one pipeline then used IGV for analysis and visual inspection of copy number variants, the other pipeline used R and the Bioconductor package ACE (absolute copy number estimation from low-coverage whole-genome sequencing data). The throughput of one flow cell achieved ~3-fold genome coverage, which was sufficient for comprehensive copy number profiling in samples with complex karyotypes.

Anna then discussed how Oxford Nanopore Direct mRNA sequencing was performed to detect translocations in the leukaemic cell line K562. From 4 μg of total RNA, ~75 ng mRNA was isolated and prepared for GridION sequencing (run time 20 hours) - a quantity "less than recommended by Nanopore but feasible for clinical samples". For analysis, minimap2 alignment was performed, followed by chimeric read selection and filtering out of low complexity reads, then BLASTn mapping of filtered reads, and finally additional filtering of aligned reads. With this pipeline, shallow transcriptome coverage was achieved in around 1 day. A wide spectrum of already known and novel structural variants was identified, including the BCR-ABL1 fusion transcript which was detected in a single run obtaining 1.2 million Direct RNA sequencing reads.

Overall, a combination of shallow whole-genome and transcriptome nanopore sequencing achieved a higher resolution and more comprehensive analysis of copy number and structural variation in AML, compared to traditional karyotyping; furthermore, nanopore sequencing achieved this in a rapid turn-around time of 1 day, as opposed to the usual 1-2 weeks.