Integrative Analysis of Gene Expression Data by RNA Sequencing for Differential Diagnosis of Acute Leukemia: Potential Application of Machine Learning

Spectrum of Genetic Mutations in Korean Pediatric Acute Lymphoblastic Leukemia

The wide application of next-generation sequencing (NGS) technologies has led to the discovery of multiple genetic alterations in pediatric acute lymphoblastic leukemia (ALL). In this work, we aimed to investigate the mutational spectrum in pediatric ALL. We employed a St. Mary’s customized NGS panel comprising 67 leukemia-related genes. Samples were collected from 139 pediatric ALL patients. Eighty-five patients (61.2%) harbored at least one mutation. In B-cell ALL, the RAS pathway is the most involved pathway, and the three most frequently mutated genes were NRAS (22.4%), KRAS (19.6%), and PTPN11 (8.4%). NRAS and PTPN11 were significantly associated with a high hyperdiploidy karyotype (p = 0.018 and p < 0.001, respectively). In T-cell ALL, the three most frequently mutated genes were NOTCH1 (37.5%), FBXW7 (16.6%), and PTEN (6.2%). Several pairs of co-occurring mutations were found: NRAS with SETD, NRAS with PTPN11 in B-cell ALL (p = 0.024 and p = 0.020, respectively), and NOTCH1 with FBXW7 in T-cell ALL (p < 0.001). The most frequent newly emerged mutation in relapsed ALL was NT5C2. We procured comprehensive genetic information regarding Korean pediatric ALL using NGS technology. Our findings strengthen the current knowledge of recurrent somatic mutations in pediatric ALL.

Predicting the influence of Circ_0059706 expression on prognosis in patients with acute myeloid leukemia using classical statistics and machine learning

Background: Various circular RNA (circRNA) molecules are abnormally expressed in acute myeloid leukemia (AML), and associated with disease occurrence and development, as well as patient prognosis. The roles of circ_0059706, a circRNA derived from ID1, in AML remain largely unclear.

Results: Here, we reported circ_0059706 expression in de novo AML and its association with prognosis. We found that circ_0059706 expression was significantly lower in AML patients than in controls (p < 0.001). Survival analysis of patients with AML divided into two groups according to high and low circ_0059706 expression showed that overall survival (OS) of patients with high circ_0059706 expression was significantly longer than that of those with low expression (p < 0.05). Further, female patients with AML and those aged >60 years old in the high circ_0059706 expression group had longer OS than male patients and those younger than 60 years. Multiple regression analysis showed that circ_0059706 was an independent factor-affecting prognosis of all patients with AML. To evaluate the prospects for application of circ_0059706 in machine learning predictions, we developed seven types of algorithm. The gradient boosting (GB) model exhibited higher performance in prediction of 1-year prognosis and 3-year prognosis, with AUROC 0.796 and 0.847. We analyzed the importance of variables and found that circ_0059706 expression level was the first important variables among all 26 factors included in the GB algorithm, suggesting the importance of circ_0059706 in prediction model. Further, overexpression of circ_0059706 inhibited cell growth and increased apoptosis of leukemia cells in vitro.

Conclusion: These results provide evidence that high expression of circ_0059706 is propitious for patient prognosis and suggest circ_0059706 as a potential new biomarker for diagnosis and prognosis evaluation in AML, with high predictive value and good prospects for application in machine learning algorithms.

APAview: A web-based platform for alternative polyadenylation analyses in hematological cancers

Background: Hematologic malignancies, such as acute promyelocytic leukemia (APL) and acute myeloid leukemia (AML), are cancers that start in blood-forming tissues and can affect the blood, bone marrow, and lymph nodes. They are often caused by genetic and molecular alterations such as mutations and gene expression changes. Alternative polyadenylation (APA) is a post-transcriptional process that regulates gene expression, and dysregulation of APA contributes to hematological malignancies. RNA-sequencing-based bioinformatic methods can identify APA sites and quantify APA usages as molecular indexes to study APA roles in disease development, diagnosis, and treatment. Unfortunately, APA data pre-processing, analysis, and visualization are time-consuming, inconsistent, and laborious. A comprehensive, user-friendly tool will greatly simplify processes for APA feature screening and mining.

Results: Here, we present APAview, a web-based platform to explore APA features in hematological cancers and perform APA statistical analysis. APAview server runs on Python3 with a Flask framework and a Jinja2 templating engine. For visualization, APAview client is built on Bootstrap and Plotly. Multimodal data, such as APA quantified by QAPA/DaPars, gene expression data, and clinical information, can be uploaded to APAview and analyzed interactively. Correlation, survival, and differential analyses among user-defined groups can be performed via the web interface. Using APAview, we explored APA features in two hematological cancers, APL and AML. APAview can also be applied to other diseases by uploading different experimental data.

Genetic Characteristics According to Subgroup of Acute Myeloid Leukemia with Myelodysplasia-Related Changes

Acute myeloid leukemia with myelodysplasia-related changes (AML-MRC) includes heterogeneous conditions such as previous history and specific cytogenetic and morphological properties. In this study, we analyze genetic aberrations using an RNA-based next-generation sequencing (NGS) panel assay in 45 patients with AML-MRC and detect 4 gene fusions of KMT2A-SEPT9, KMT2A-ELL, NUP98-NSD1, and RUNX1-USP42 and 81 somatic mutations. Overall, all patients had genetic aberrations comprising of not only cytogenetic changes, but also gene fusions and mutations. We also demonstrated several characteristic genetic mutations according to the AML-MRC subgroup. TP53 was the most commonly mutated gene (n = 11, 24%) and all were found in the AML-MRC subgroup with myelodysplastic syndrome-defining cytogenetic abnormalities (AML-MRC-C) (p = 0.002). These patients showed extremely poor overall survival not only in AML-MRC, but also within the AML-MRC-C subgroup. The ASXL1 (n = 9, 20%) and SRSF2 (n = 7, 16%) mutations were associated with the AML-MRC subgroup with >50% dysplasia in at least two lineages (AML-MRC-M) and were frequently co-mutated (55%, 6/11, p < 0.001). Both mutations could be used as surrogate markers to diagnose AML-MRC, especially when the assessment of multilineage dysplasia was difficult. IDH1/IDH2 (n = 13, 29%) were most commonly mutated in AML-MRC, followed by CEBPA (n = 5, 11%), PTPN11 (n = 5, 11%), FLT3 (n = 4, 9%), IDH1 (n = 4, 9%), and RUNX1 (n = 4, 9%). These mutations were not limited in any AML-MRC subgroup and could have more significance as a risk factor or susceptibility marker for target therapy in not only AML-MRC, but also other AML categories.

Multi-Cohort Transcriptomic Subtyping of B-Cell Acute Lymphoblastic Leukemia

RNA sequencing provides a snapshot of the functional consequences of genomic lesions that drive acute lymphoblastic leukemia (ALL). The aims of this study were to elucidate diagnostic associations (via machine learning) between mRNA-seq profiles, independently verify ALL lesions and develop easy-to-interpret transcriptome-wide biomarkers for ALL subtyping in the clinical setting. A training dataset of 1279 ALL patients from six North American cohorts was used for developing machine learning models. Results were validated in 767 patients from Australia with a quality control dataset across 31 tissues from 1160 non-ALL donors. A novel batch correction method was introduced and applied to adjust for cohort differences. Out of 18,503 genes with usable expression, 11,830 (64%) were confounded by cohort effects and excluded. Six ALL subtypes (ETV6::RUNX1, KMT2A, DUX4, PAX5 P80R, TCF3::PBX1, ZNF384) that covered 32% of patients were robustly detected by mRNA-seq (positive predictive value ≥ 87%). Five other frequent subtypes (CRLF2, hypodiploid, hyperdiploid, PAX5 alterations and Ph-positive) were distinguishable in 40% of patients at lower accuracy (52% ≤ positive predictive value ≤ 73%). Based on these findings, we introduce the Allspice R package to predict ALL subtypes and driver genes from unadjusted mRNA-seq read counts as encountered in real-world settings. Two examples of Allspice applied to previously unseen ALL patient samples with atypical lesions are included.