The dendrogram identifies six modules based on the similarity of expression patterns among their genes. The modules are denoted by various colors (e.g., MEyellow, MEblue, MEturquoise, etc.), and their association with the population groups is indicated by numerical values and significance levels (Fig. 4). Based on the analysis of the link between modules and traits, a significant positive or negative association with either group is observed. A positive number signifies that the genes within this module have more expression in AAM compared to EAM. A negative number signifies that the genes within the module have lower levels of expression in AAM compared to EAM.

This implies that the genes inside these modules exhibit distinct behavior among different population groups, indicating the presence of biological processes or mechanisms in cancer that are affected by these difference. These modules are crucial for comprehending the distinct manifestation or progression of cancer in various populations.

The blue and yellow modules exhibit a strong correlation with the prostate cancer pathway, suggesting that the genes within these modules might contribute to the development and progression of prostate cancer. By prioritizing the modules that are enhanced in prostate cancer pathways, we may potentially discover crucial genes and molecular interactions that are unique to the disease. These findings could serve as possible biomarkers for diagnosis or targets for treatment. Conducting functional tests to validate the functions of these genes and pathways in prostate cancer is crucial.

A total of 104 genes were discovered using LASSO. Using CFS with a correlation threshold of 1, a group of 10 genes was subsequently found. Notably, six genes were found to be common between these two approaches. The precision rate of the six shared genes achieved a level of 73%. The identification of these six genes, through the combined utilization of LASSO and CFS techniques, signifies a momentous advancement in comprehending the essential genetic elements that contribute to prostate cancer. The presence of these often-found genes suggests a significant connection to the studied condition, as they are supported by both research methods. Additional investigation and examination of these genes may provide a crucial understanding of their functional roles, pathways, and their significance in relation to the causes, prognosis, or therapy approaches for prostate cancer. Furthermore, exploring the regulatory networks or interactions involving these genes could reveal new insights into comprehending the process of the disease.

Adenine phosphoribosyltransferase (APRT) is a metabolic enzyme that participates in the production of polyamines, which are essential for the rapid growth of cancer cells. APRT (Adenine Phosphoribosyltransferase) has the potential to be a target for cancer treatment, as suppressing the APRT gene has harmful effects on leukemia cell lines [18].

CCL2 is involved in the onset and advancement of several types of malignancies. It can stimulate the growth and multiplication of tumor cells through various mechanisms and facilitate the migration of cancer cells. Additionally, it can attract cells that inhibit the immune system to the surrounding environment of the tumor, thereby promoting the progression of cancer [19]. CCL2 is the most potent chemoattractant in the tumor microenvironment, responsible for attracting macrophages and initiating inflammation. It exerts chemotactic effects on neighboring host cells inside the tumor microenvironment and collaboratively influences their differentiation with other cytokines. Nevertheless, the presence of CCL2 in tumor patients leads to a detrimental impact on their prognosis, as it leads to the buildup of cell subtypes that suppress the immune system [20]. In addition, CCL2 attracts immune cells, specifically monocytes and macrophages, which subsequently transform into immunosuppressive myeloid-derived suppressor cells (MDSCs) and M2 macrophages. This recruitment worsens the immunosuppressive tumor microenvironment and undermines the effectiveness of treatment. In their 2021 study, Liu and colleagues discovered that CCL2 is the primary mediator released by tumor-associated adipocytes into the surrounding extracellular environment. They also developed a protein trap that effectively binds to CCL2 with strong affinity and specificity, allowing for the manipulation of CCL2-mediated immune responses. This approach demonstrated improved treatment effectiveness and significant suppression of tumor development [21].

BEX2 and its homolog BEX1 have a strong correlation in their expression and are members of a cluster that is enriched with genes involved in the ER response and apoptosis. The gene BEX2 has been recognized as being expressed at higher levels in a specific group of breast tumors that have estrogen receptors (ER). Additionally, it has been linked to better results following treatment with tamoxifen [22]. Nevertheless, there is a lack of explicit data about the involvement of BEXL1 in cancer.

MGC26963, alternatively referred to as Sphingomyelin synthase 2 (SGMS2), is a genetic element that has been associated with multiple forms of cancer. Research has demonstrated a significant association between the expression of SGMS2 mRNA and the presence of tumor-associated macrophages (TAMs), as well as a negative impact on the prognosis of patients with pancreatic ductal adenocarcinoma (PDAC) [23]. High levels of M2-polarized macrophages in the original tumor of triple-negative breast cancer (TNBC) are linked to a dismal prognosis. Blocking SGMS2 or genetically eliminating its expression decreases the M2 polarization of tumor-associated macrophages and hinders the advancement of tumors in triple-negative breast cancer (TNBC) [24]. Ovarian cancer exhibits a unique upregulation of SMS2, which actively promotes the migration, development, and survival of cancer cells. Suppression of SMS2 by depletion or inhibition hinders the migration, development, and survival of ovarian cancer cells [25]. SGMS2 enhances the growth and spread of cancer cells in breast cancer by utilizing a mechanism connected with ceramide and activating the TGF-β/Smad signaling pathway [26]. Using a mouse model, the absence of SMS2 hinders the development of the tumor microenvironment and prevents the entry of cancer cells [27].

PLAU, the urokinase-type plasminogen activator, exerts a substantial influence on the advancement of cancer. It facilitates cell growth, movement, attachment, and various other activities using the proteolytic system, intracellular signal transmission, and chemokine activation [28]. Increased PLAU expression is linked to heightened aggressive characteristics, stromal score, and immune suppression in pancreatic ductal adenocarcinoma (PDAC) [29]. PLAU is additionally linked to the movement and infiltration of cells and is controlled by the transcription factor YY1 in cervical cancer [30]. In addition, PLAU, sometimes referred to as a urokinase-type plasminogen activator (uPA), stimulates the movement, infiltration, and multiplication of colorectal cancer cells through the Src/ERK pathway [31]. Hence, directing efforts towards PLAU could potentially yield diagnostic, prognostic, and therapeutic benefits in many cancer types [32].

Remarkably, the analysis of both the LASSO and CFS approaches has led to the detection of six probes, with four of them located within the yellow module. The significance of the yellow module in the context of prostate cancer research is emphasized by this association. Furthermore, it has been noted that the genes in the yellow module demonstrate elevated levels of expression in AAM in comparison to EAM. This indicates a possible gene expression pattern that is specific to certain populations, which could have significant ramifications for the susceptibility to diseases and prognosis.

We employed Weighted WGCNA to investigate the genetic characteristics of prostate cancer across various population groups. The modules revealed in the investigation of the link between modules and traits reveal a clear gene expression pattern that is associated with different population backgrounds. These findings indicate that population factor (AAM vs. EAM) have a certain degree of influence on the genetic basis of prostate cancer. The genes found by LASSO (Linkage Analysis of Sequence Outliers) and CFS (Correlation-based Feature Selection) provide promising targets for comprehending the molecular mechanisms underlying these differences.

The discovery of genetically related modules in prostate cancer that are associated with race is consistent with prior studies that have demonstrated variances in genes among different races. Research has indicated that African-American men have a greater occurrence and severity of prostate cancer, possibly due to the varying activity of specific genes. Our research emphasizes particular gene clusters and genes that may play a crucial role in these variances.

Further work is necessary for the noteworthy modules and genes. Their prominent position in the gene networks implies that they could be crucial catalysts for the biological processes linked to disparities in prostate cancer. The Gene Ontology study offered further context by establishing connections between these genes and distinct cellular processes and molecular activities, thus enhancing our overall comprehension of their potential influence.

These discoveries create opportunities for more focused genomic investigations and potentially individualized therapeutic approaches. Gaining insight into the genetic determinants responsible for disparities in prostate cancer among different population groups may result in the development of more efficient screening, diagnosis, and treatment procedures that are customized for distinct populations. Moreover, including these genetic markers in clinical trials has the potential to advance the creation of treatments that are very efficient in many populations. Although our study offers valuable insights, it does have limits. Dependence on publicly accessible microarray datasets can lead to biases and limit the generalizability of the findings to all population groups. Subsequent investigations should prioritize the verification of these discoveries via clinical trials and broaden the scope of the analysis to encompass a more extensive range of genetic information. Incorporating environmental and lifestyle factors could provide a comprehensive perspective on the underlying causes of differences in prostate cancer.

Ultimately, our study emphasizes the significance of taking population characteristics into account when conducting a genetic analysis of prostate cancer. The identified gene modules and genes offer a fundamental comprehension of the molecular variations that could potentially contribute to the reported discrepancies in prostate cancer occurrence and advancement among various population groups. This research not only contributes to current knowledge but also emphasizes the necessity for individualized approaches in cancer therapy and care.

Our study deepens significant differences in gene expression patterns of prostate cancer between African American men (AAM) and European American men (EAM). These findings may be essential to develop personalized diagnosis resulted in more effective therapeutic strategies. The identification of potential biomarkers such as APRT, CCL2, BEX2, MGC26963, and PLAU through specific gene modules and key genes could enhance our perception of prostate cancer's molecular mechanisms and targeted treatments. However, several limitations could introduce potential biases. The public microarray gene expression profile (GSE41967) from a single geographic location and timeframe may not be representative of other populations or current clinical settings. The lack of information on metastatic disease and tumor characteristics, as well as the focus exclusively on AAM and EAM groups, limits the broader applicability of the results. Lifestyle factors, environmental exposures, and socioeconomic status may significantly affect cancer risk and progression, however, they were not considered in this study. Furthermore, microarray technology, while robust, has limitations compared to newer sequencing technologies that may potentially affect the resolution and sensitivity of gene expression differences. As well, Analytical methods such as WGCNA, LASSO regression, and CFS, have inherent biases related to their algorithmic assumptions. Reliance on publicly available datasets may introduce biases associated with sample selection and original study designs. Future studies is suggested to incorporate more diverse populations, consider environmental factors and socioeconomic, use advanced genomic technologies, and validate findings with independent datasets to enhance the robustness and applicability of the results.