Covering Hierarchical Dirichlet Mixture Models on binary data to enhance genomic stratifications in onco-hematology

Onco-hematological studies are increasingly adopting statistical mixture models to support the advancement of the genomically-driven classification systems for blood cancer. Targeting enhanced patients stratification based on the sole role of molecular biology attracted much interest and contributes to bring personalized medicine closer to reality. In onco-hematology, Hierarchical Dirichlet Mixture Models (HDMM) have become one of the preferred method to cluster the genomics data, that include the presence or absence of gene mutations and cytogenetics anomalies, into components. This work unfolds the standard workflow used in onco-hematology to improve patient stratification and proposes alternative approaches to characterize the components and to assign patient to them, as they are crucial tasks usually supported by a priori clinical knowledge. We propose (a) to compute the parameters of the multinomial components of the HDMM or (b) to estimate the parameters of the HDMM components as if they were Multivariate Fisher's Non-Central Hypergeometric (MFNCH) distributions. Then, our approach to perform patients assignments to the HDMM components is designed to essentially determine for each patient its most likely component. We show on simulated data that the patients assignment using the MFNCH-based approach can be superior, if not comparable, to using the multinomial-based approach. Lastly, we illustrate on real Acute Myeloid Leukemia data how the utilization of MFNCH-based approach emerges as a good trade-off between the rigorous multinomial-based characterization of the HDMM components and the common refinement of them based on a priori clinical knowledge.

AML with complex karyotype: extreme genomic complexity revealed by combined long-read sequencing and Hi-C technology

Acute myeloid leukemia with complex karyotype (CK-AML) is associated with poor prognosis, which is only in part explained by underlying TP53 mutations. Especially in the presence of complex chromosomal rearrangements, such as chromothripsis, the outcome of CK-AML is dismal. However, this degree of complexity of genomic rearrangements contributes to the leukemogenic phenotype and treatment resistance of CK-AML remains largely unknown. Applying an integrative workflow for the detection of structural variants (SVs) based on Oxford Nanopore (ONT) genomic DNA long-read sequencing (gDNA-LRS) and high-throughput chromosome confirmation capture (Hi-C) in a well-defined cohort of CK-AML identified regions with an extreme density of SVs. These rearrangements consisted to a large degree of focal amplifications enriched in the proximity of mammalian-wide interspersed repeat elements, which often result in oncogenic fusion transcripts, such as USP7::MVD, or the deregulation of oncogenic driver genes as confirmed by RNA-seq and ONT direct complementary DNA sequencing. We termed this novel phenomenon chromocataclysm. Thus, our integrative SV detection workflow combing gDNA-LRS and Hi-C enables to unravel complex genomic rearrangements at a very high resolution in regions hard to analyze by conventional sequencing technology, thereby providing an important tool to identify novel important drivers underlying cancer with complex karyotypic changes.

Clonal evolution patterns in acute myeloid leukemia with NPM1 mutation

Mutations in the nucleophosmin 1 (NPM1) gene are considered founder mutations in the pathogenesis of acute myeloid leukemia (AML). To characterize the genetic composition of NPM1 mutated (NPM1^mut) AML, we assess mutation status of five recurrently mutated oncogenes in 129 paired NPM1^mut samples obtained at diagnosis and relapse. We find a substantial shift in the genetic pattern from diagnosis to relapse including NPM1^mut loss (n = 11). To better understand these NPM1^mut loss cases, we perform whole exome sequencing (WES) and RNA-Seq. At the time of relapse, NPM1^mut loss patients (pts) feature distinct mutational patterns that share almost no somatic mutation with the corresponding diagnosis sample and impact different signaling pathways. In contrast, profiles of pts with persistent NPM1^mut are reflected by a high overlap of mutations between diagnosis and relapse. Our findings confirm that relapse often originates from persistent leukemic clones, though NPM1^mut loss cases suggest a second "de novo" or treatment-associated AML (tAML) as alternative cause of relapse.