Optimizing HAMLET for Advanced RNA-Seq Diagnostics in Acute Myeloid Leukemia
Acute Myeloid Leukemia (AML) is a fast growing cancer of the myeloid
lineage which can be caused by a wide variety of genomic aberrations,
varying from small single nucleotide variants to large scale chromosomal
rearrangements. Due to this diversity, uncovering the causal genetic
changes for patients diagnosed with AML can be challenging and time
consuming. We have developed HAMLET, an automated pipeline to call
genetic aberrations for diagnosis, prognosis and treatment of AML.
To capture all genetic aberrations relevant for diagnosis of AML, we
used whole transcriptome RNA sequencing to screen all expressed genes
with a single technology. After sequencing RNA extracted from peripheral
blood or bone marrow, HAMLET was used to call variants in 46 genes
relevant for diagnosis of AML, and report fusion events involving a
further 46 genes. HAMLET is particularly sensitive in detecting large
tandem duplications, including tandem duplications in FLT3 and KMT2A.
Special care has been taken to ensure that it is easy for different
centers to customize HAMLET to their own needs, for example by modifying
the filter criteria for variants and gene fusions.
HAMLET has been in use for routine diagnostics in the LUMC since 2020,
and has been used to diagnose over 300 primary and relapse AML patients
under the ISO accreditation (ISO 15189:2012) of the hematologic
laboratory (LSH LUMC).
The results from HAMLET have been validated against existing
technologies for variants, fusion genes and tandem duplications in 393
samples. For 50 samples enriched for cases diagnosed as
myelodysplasia-related AML by RNA sequencing at the LUMC, results have
been validated by targeted NGS using single-molecule Molecular Inversion
Probes at the RadboudUMC (smMIPs). Results showed good concordance (98%)
between RNA sequencing and smMIPs for (likely) pathogenic variants,
indicating that using RNA for variant calling is as reliable as targeted
DNA sequencing.
Finally, we extended our HAMLET pipeline by integration novel algorithms
to predict the AML subtype and the cell type composition using gene
expression profiling.
RNA sequencing is an efficient approach to simultaneously capture both
the genetic abnormalities as well as gene expression signatures which
are relevant in diagnosing AML. In cases where there are no clear driver
mutations for AML, the additional information from RNA sequencing data
can be used to improve diagnostics or suggest treatment options. We have
automated the process of analyzing RNA data using HAMLET, a highly
configurable, easy to install and free to use RNA sequencing analysis
pipeline to aid in the diagnosis of AML.