Abstract
Cutaneous melanoma (CM) is the most fatal type of skin cancer with a high potency of metastasis, yet the treatment for metastatic melanoma remains limited. In this study, we are devoted to addressing the prognostic value and underlying mechanism of DNA damage repair-related genes in CM. We utilized integrated bioinformatic approaches and machine learning models to identify a cluster of convergently expressed DNA damage repair-related genes in melanoma. With multivariate Cox regression, SMARCA4 (also known as BRG1) was identified as an independent prognostic marker for melanoma patients. Yet the expression of SMARCA4 is not altered with the pathological staging or the metastasis condition. SMARCA4 is an essential ATPase subunit of the mammalian SWI/SNF complex. Mechanistically, we demonstrated that SMARCA4 could resolve DNA replication stress and guarantee the proliferation of melanoma cells. Furthermore, we predicted the binding of different transcription factors on the SMARCA4 promoter and unveiled the modulated expression of SMARCA4 by SOX10 in melanoma. Together, we performed integrated approaches to identify SMARCA4 as a promising prognostic marker for melanoma, which was transcriptionally regulated by SOX10 and promoted melanoma cell proliferation by ameliorating DNA replication stress.