Projects

In silico analysis of next generation sequencing of Melanoma, Oral Cavity, and Acute Myeloid Leukemia samples for neoantigen prediction, TCR and BCR rearrangement identification, and tumor markers evaluation. Immunological response against transformed cells are antigen specific and capable of controlling tumor development. However, in some cases, tumors escape immunologic surveillance by losing the more immunogenic variants. This impacts the neoantigen burden, the amount of new antigens generated by coding change somatic mutations. Analyzing these cancer types yields a more general landscape, without preventing us to explore each disease’s particularity. Melanoma possess one of the highest mutation burdens and neoantigen generation which may favour immune responses against tumours. Recent therapies such as immunologic checkpoint inhibitors show marked clinical benefits in a fraction of melanoma patients. Myeloproliferative neoplasm (MPN) are diseases impacting the myeloid compartment in which mutations in driver genes and/or epigenetic modulators lead to expansion of mature myeloid derived populations. Such diseases can eventually progress to secondary acute myeloid leukemia (sAML), having a dismal prognosis. Limited data has been generated regarding their neoantigen profile and how these profiles are modulated during the MPN→sAML transition.

In silico evaluation of pathways modulated by the ZNF429 transcription factor in ovarian cancer. Among gynecologial cancers, ovarian cancer is responsible for approximately a third of all cases. Since patients don't present symptoms until the most advanced stages of this disease, they usually have late diagnosis. The existence of many molecular subtypes with distinct molecular patterns imply different levels of chemosensibility, metastasis and prognosis, point to the need to identify markers that facilitate the rapid identification of the tumor molecular pattern and patient prognosis. ZNF429 has been identified as an important gene, with a positive prognostic impact on the survival rate of patients with ovarian cancer. ZNF429 is part of a family of transcriptional regulators with a KRAB-like domain in their constitution, a strong reporter of gene transcription, which are important in cell regulation. Despite the lack of knowledge about ZNF429 cellular functionality, some studies point to its importance in the tumor environment. This makes the study of its expression patterns, the understanding of its prognostic power and the identification of genes regulated by it of great relevance due to its potential of helping us better understand the progression of the disease and identify possible new therapeutic targets.

Identification of prognostic markers in ovarian cancer using machine learning. Ovarian cancer is one of the neoplasms with the highest incidence among women worldwide, with significantly high mortality. The vast majority of patients are diagnosed in advanced stages of the disease, as the early stages show nonspecific symptoms and inaccurate diagnoses. Currently, there are not many biomarkers available clinically to help diagnose or predict your prognosis because of contradictory diagnostic accuracy. In this sense, the application of multi-omic integration approaches combined with machine learning techniques is promising, not only to better understand the cancer prognosis but to identify more effective prognostic biomarkers related to ovarian cancer. Therefore, our goal is to build a multi-omic predictor of prognosis for patients diagnosed with ovarian cancer, as well as to identify new biomarkers using machine learning techniques based on clinical-pathological and molecular data from samples of patients diagnosed with cancer obtained from the TCGA database (The Cancer Genome Atlas).

Characterization of the interaction networks between circRNA-mirRNA-mRNA in ovarian tumors with different tumor microenvironment profiles. Ovarian cancer is the most common cause of death associated with gynecological cancer and although there are treatment approaches, further studies are clearly needed to better understand the prognosis and develop more effective therapeutic options. Among the various study approaches, data increasingly indicate that the progression of cancer is associated with the presence or absence of certain cellular subtypes in the tumor microenvironment (TME) that promote mechanisms such as tumor escape and promotion of metastasis through communication with neoplastic cells. With this in mind, this project aims to characterize the microenvironment of ovarian tumors based on gene signatures, and classify them in groups according to the TME profile and prognosis. Then, to characterize the interaction of miRNA and circRNA - identified as important markers of the communication of the TME with the tumor - that are differentially expressed among the groups, to obtain the important biological pathways that are related to them and to understand the association between survival and the expression of these RNAs.

Identification of potential therapeutic targets in colorectal cancer. Colorectal cancer is one of the most incident carcinomas in the world, being the fourth in the world and the second in Brazil, according to the most recent estimates. This cancer has a high heterogeneity, resulting in tumors with varied prognosis and response to treatment. Recent research has been carried out in order to better understand this heterogeneity, represented by molecular subtypes. Each subtype has distinct expression and mutation patterns as well as different prognosis. This indicates a potential for the development of specific treatment strategies. Based on this, the aim of this project is to identify, using the molecular subtypes' expression patterns, potential new subtype-specific therapeutic targets. For this, the patterns of gene expression of the molecular subtypes of colorectal cancer will be used for the construction of interaction networks, with priority being given to the most important genes for the validation of drugs targeting them.

Evaluation of epigenetic and miRNA profiles of melanomas with BRAF, NRAS, NF1, and Triple Wild-type. Cutaneous melanoma is one of the leading causes of death related to skin cancer. Most melanoma tumors present recurrent mutations in BRAF, RAS, NF1 and KIT genes, charactering distinct tumor subtypes. Another characteristic of this disease is the intense inflammatory infiltrate found in the tumor micro-environment accompanied by several mechanisms of immune escape used by the tumor. Changes in the epigenetic and miRNAs expression profile are found in many types of cancer and can affect tumorigenesis through regulating tumor-associated stromal components including fibroblasts and immune cells. Epigenetic alterations may also drive immune population infiltrate shifts. This project aims to evaluate the influence of the aberrant miRNA expression and methylation alterations in the inflammatory profile on the four different melanoma subtypes and its impact on overall survival.

Evaluation of lncRNAs as potential prognostic biomarkers in metastatic melanoma using machine learning. Melanoma, despite having approximately 5% of incidence among skin neoplasms, is the most lethal of them. Although immunotherapies involving immune checkpoint blockade such as CTLA-4 and PD-1 have shown promising results, there is a great variation in response to treatment, mainly related to genomic heterogeneity and immune infiltrate in patients. Therefore, prognostic biomarkers play a critical role in understanding disease progression and optimizing treatment and patient survival. Long non-coding RNAs (LncRNA) have already been described as prognostic biomarkers of several types of cancer, including cutaneous melanomas. In this sense, our work aims to investigate the role of lncRNAs as prognostic biomarkers specifically in melanoma patients who have metastasized, and to develop a model for predicting patient survival based on the expression profile of lncRNAs. We use machine learning (ML) techniques such as supervised machine learning algorithms on clinical-pathological data and gene expression of lncRNA from samples of patients diagnosed with metastatic melanoma obtained from the database TCGA (The Cancer Genome Atlas).

Characterization of the molecular mechanism responsible for FLT3 gene overexpression in Acute Leukemias. FLT3 overexpression is a recurrent alteration in acute leukemias (ALs), which can lead to the constitutive activation of its tyrosine kinase receptor by a ligand-independent mechanism, resulting in increased cell proliferation, reduced apoptosis, and inhibition of cell differentiation. Over time, many studies have associated this alteration with the presence of activating mutations in FLT3, which is considered the main cause of transcriptional dysregulation of this gene. However, many patients with FLT3 overexpression do not have these mutations, suggesting that there are other mechanisms, still unknown, responsible for this alteration. In this context, the integration of several omics data has enabled the identification of enhancer regions associated with molecular alterations leading to deregulation of the expression of several proto-oncogenes. Thus, we hypothesized that the activation mechanism through the neomorphic enhancers or super-enhancers may be responsible for FLT3 overexpression in a portion of AL cases. Therefore, we aim to search for new molecular mechanisms responsible for FLT3 gene overexpression in ALs.