ABSTRACT
The OmicsFootPrint framework addresses the need for advanced multi-omics data analysis methodologies by transforming data into intuitive two-dimensional circular images and facilitating the interpretation of complex diseases. Utilizing Deep Neural Networks and incorporating the SHapley Additive exPlanations (SHAP) algorithm, the framework enhances model interpretability. Tested with The Cancer Genome Atlas (TCGA) data, OmicsFootPrint effectively classified lung and breast cancer subtypes, achieving high Area Under Curve (AUC) scores - 0.98±0.02 for lung cancer subtype differentiation, 0.83±0.07 for breast cancer PAM50 subtypes, and successfully distinguishe between invasive lobular and ductal carcinomas in breast cancer, showcasing its robustness. It also demonstrated notable performance in predicting drug responses in cancer cell lines, with a median AUC of 0.74, surpassing existing algorithms. Furthermore, its effectiveness persists even with reduced training sample sizes. OmicsFootPrint marks an enhancement in multi-omics research, offering a novel, efficient, and interpretable approach that contributes to a deeper understanding of disease mechanisms.
ABSTRACT
BACKGROUND: The most common male malignancy in the United States is prostate cancer; however its rate of occurrence varies significantly among ethnic groups. In a previous cDNA microarray study on CaP tumors from African American (AA) and Caucasian (CA) patients, we identified 97 candidate genes that exhibited opposite gene expression polarity with respect to race groups; genes up-regulated in AA were simultaneously down-regulated in CA. PURPOSE: The purpose of this study was to narrow the 97 member gene list, to a smaller number of genes in order to focus studies on a limited number of genes/SNPs that might explain prostate cancer disparity in African Americans. METHODS: We performed genotype-phenotype, SNP and expression transcript levels correlations using HapMap Yoruba population with 85 of our 97 prostate candidate genes using SCAN database. RESULTS: Findings revealed an association of SNPs surrounding ABCD3 gene with basal gene expression of RanGAP1 is important in prostate tumors in AA. Hence, to confirm our results in clinical biospecimen, we monitored expression of ABCD3 in a novel panel of African American and Caucasian prostate cancer paired cell lines. The LNCaP, C4-2B showed 2-fold increase; MDA-2PC-2B cell line, derived from AA, showed highest fold-change, 10-fold. The EGFR over expressing DU-145 WT cell line exhibited a 4-fold increase in expression relative to non transfected DU-145 prostate cell lines. Furthermore, Ingenuity Network analysis implicated our AA prostate candidate genes are involved in three network hubs, ERK, MapK and NFkB pathways. CONCLUSIONS: Taken together, these findings are intriguing because other members of the ABC gene family, namely, ABCC3, ABCD1, and ABCD2 have been shown to confer chemoresistance in certain cancer types. Equally important, is the fact that activation of the MapK/ERK pathway via EGFR stimulation is vital for increased transcription of numerous cancer related genes. It is especially noteworthy that overexpression of EGFR has been widely observed in AA prostate tumors. Collectively our findings lead us to think that a novel signaling cascade, through which increased aggressiveness and chemoresistance is achieved, may explain prostate cancer health disparity in AA males and the nature of aggressive CaP tumors in general.