Chronic myeloid leukemia (CML) is a type of cancer that is classified into three phases:Chronic, Accelerated, and Blast crisis, while the myeloid blast crisis (MBC) phase is resistant totreatment [۱]. Despite evidence in the BCR-ABL fusion gene as the most significant geneticabnormality and a key for the initiation and supporting the continuation of CML, there are fascinatingproofs that
CML harbors many other genetic alterations in oncogenes as well as tumor suppressor genesthat contribute to the progression of this malignancy. Some of these changes are drivers that map tosignaling pathways and control cell growth, cell cycle progression, and apoptosis. While the others actas cancerpredisposing variants, most of them are germline. The affected genes may work together inconnected networks as modules and play a role in the pathogenesis of MBC-CML. In this study, wetried to find potential disease modules, as a set of genes that enrich the novel variants in cases withMBC-CML compared to controls which were Imatinib-responsive CML. To this end, we utilized theMAGI algorithm [۲], which merges different omics data including gene co-expression network,ProteinProtein interaction, and genetic variant data in case and control groups. MAGI has been used forneurodevelopmental diseases such as autism and Intellectual disability. But for the first time, wepropose using this algorithm for cancer diseases. To this end, we used HPRD data as PPI network input,and we constructed a gene co-expression network for
CML patients based on the GSE۴۲۵۱۹ microarraydataset [۳]. We also used European Genome-Phenome Archive (EGA) data with accessionEGAS۰۰۰۰۱۰۰۳۰۷۱ as our case and control samples [۴]. We selected ۱۸ samples with the MBC-CMLcase group and ۴۴ samples with the
CML control group. We analyzed the fastq files based on GRCh۳۸genome assembly using BWAMEM as the aligner, and Mutect۲ as the variant caller. We used theMutect۲ algorithm from the GATK۴ toolkit to call somatic and germline variants. Then Several toolswere used to annotate and interpret variants including CGI, Varsome, InterVar, CancerVar, and Franklin.The annotated VCF files were filtered to find Tier ۱ and Tier ۲ variants according to AMP classificationand Tier ۳ variants which were pathogenic (P) based on ACMG classification. Finally, we created atable of important gene variants for case and control groups. Then the
MAGI algorithm was run to findthe modules including novel genes based on PPI and co-expression network. The Module with the bestscore was selected as the best module and enriched with the EnrichR web tool. The Enriched termswere highly associated with
CML and cancer and as a result, the novel genes can be potentiallyassociated genes with
CML resistance to the treatment.