Beatriz Stransky

Bioinformatics construction of the human cell surfaceome

Cell surface proteins are excellent targets for diagnostic and therapeutic interventions. By using bioinformatics tools, we generated a catalog of 3,702 transmembrane proteins located at the surface of human cells (human cell surfaceome). We explored the genetic diversity of the human cell surfaceome at different levels, including the distribution of polymorphisms, conservation among eukaryotic species, and patterns of gene expression. By integrating expression information from a variety of sources, we were able to identify surfaceome genes with a restricted expression in normal tissues and/or differential expression in tumors, important characteristics for putative tumor targets. A high-throughput and efficient quantitative real-time PCR approach was used to validate 593 surfaceome genes selected on the basis of their expression pattern in normal and tumor samples. A number of candidates were identified as potential diagnostic and therapeutic targets for colorectal tumors and glioblastoma. Several candidate genes were also identified as coding for cell surface cancer/testis antigens. The human cell surfaceome will serve as a reference for further studies aimed at characterizing tumor targets at the surface of human cells.

MODELING CANCER: INTEGRATION OF "OMICS" INFORMATION IN DYNAMIC SYSTEMS

The last 10 years have seen the rise of many technologies that produce an unprecedented amount of genome-scale data from many organisms. Although the research community has been successful in exploring these data, many challenges still persist. One of them is the effective integration of such data sets directly into approaches based on mathematical modeling of biological systems. Applications in cancer are a good example. The bridge between information and modeling in cancer can be achieved by two major types of complementary strategies. First, there is a bottom-up approach, in which data generates information about structure and relationship between components of a given system. In addition, there is a top-down approach, where cybernetic and systems-theoretical knowledge are used to create models that describe mechanisms and dynamics of the system. These approaches can also be linked to yield multi-scale models combining detailed mechanism and wide biological scope. Here we give an overall picture of this field and discuss possible strategies to approach the major challenges ahead.

Modeling tumor evolutionary dynamics

Tumorigenesis can be seen as an evolutionary process, in which the transformation of a normal cell into a tumor cell involves a number of limiting genetic and epigenetic events, occurring in a series of discrete stages. However, not all mutations in a cell are directly involved in cancer development and it is likely that most of them (passenger mutations) do not contribute in any way to tumorigenesis. Moreover, the process of tumor evolution is punctuated by selection of advantageous (driver) mutations and clonal expansions. Regarding these driver mutations, it is uncertain how many limiting events are required and/or sufficient to promote a tumorigenic process or what are the values associated with the adaptive advantage of different driver mutations. In spite of the availability of high-quality cancer data, several assumptions about the mechanistic process of cancer initiation and development remain largely untested, both mathematically and statistically. Here we review the development of recent mathematical/computational models and discuss their impact in the field of tumor biology.

Accelerating gene regulatory networks inference through GPU/CUDA programming

Gene regulatory networks (GRN) inference is an important bioinformatics problem in which the gene interactions need to be deduced from gene expression data, such as microarray data. Feature selection methods can be applied to this problem. A feature selection technique is composed by two parts: a search algorithm and a criterion function. Among the search algorithms already proposed, there is the exhaustive search where the best feature subset is returned, although its computational complexity is unfeasible in almost all situations. The objective of this work is the development of a low cost parallel solution based on GPU architectures for exhaustive search with a viable cost-benefit. CUDA™ is a general purpose parallel architecture with a new parallel programming model allowing that the NVIDIA® GPUs solve complex problems in an efficient way. We developed a parallel algorithm for GRN inference based on the GPU/CUDA and encouraging speedups (60x) were achieved when assuming that each target gene has two predictors. The idea behind the proposed method can be applied considering three or more predictors for each target gene as well.

Genome-wide identification of cancer/testis genes and their association with prognosis in a pan-cancer analysis

Cancer/testis (CT) genes are excellent candidates for cancer immunotherapies because of their restrict expression in normal tissues and the capacity to elicit an immune response when expressed in tumor cells. In this study, we provide a genome-wide screen for CT genes with the identification of 745 putative CT genes. Comparison with a set of known CT genes shows that 201 new CT genes were identified. Integration of gene expression and clinical data led us to identify dozens of CT genes associated with either good or poor prognosis. For the CT genes related to good prognosis, we show that there is a direct relationship between CT gene expression and a signal for CD8+ cells infiltration for some tumor types, especially melanoma.

Bioinformatics Analysis of the Human Surfaceome Reveals New Targets for a Variety of Tumor Types

It is estimated that 10 to 20% of all genes in the human genome encode cell surface proteins and due to their subcellular localization these proteins represent excellent targets for cancer diagnosis and therapeutics. Therefore, a precise characterization of the surfaceome set in different types of tumor is needed. Using TCGA data from 15 different tumor types and a new method to identify cancer genes, the S-score, we identified several potential therapeutic targets within the surfaceome set. This allowed us to expand a previous analysis from us and provided a clear characterization of the human surfaceome in the tumor landscape. Moreover, we present evidence that a three-gene set—WNT5A, CNGA2, and IGSF9B—can be used as a signature associated with shorter survival in breast cancer patients. The data made available here will help the community to develop more efficient diagnostic and therapeutic tools for a variety of tumor types.

Insights into gliomagenesis: systems biology unravels key pathways

Technological advances have enabled a better characterization of all the genetic alterations in tumors. A picture that emerges is that tumor cells are much more genetically heterogeneous than originally expected. Thus, a critical issue in cancer genomics is the identification of the genetic alterations that drive the genesis of a tumor. Recently, a systems biology approach has been used to characterize such alterations and find associations between them and the process of gliomagenesis. Here, we discuss some implications of this strategy for the development of new therapeutic and diagnostic protocols for cancer.

Modelling the Effects of Genetic Changes in Tumour Progression

Tumours can be considered a set of cells that accumulate genetic and epigenetic alterations. According to the Multi-stage Hit theory, the transformation of a normal into a tumour cell involves a number of limiting events that occur in a number of discrete stages (driver mutations). However, not all mutations that occur in the cell are directly involved in the development of cancer and some probably do not contribute in any way (passenger mutations). Moreover, the process of tumour evolution is punctuated by selection of advantageous mutations and clonal expansions. Actually, it is not known how many limiting-events, i.e., how many driver mutations are necessary or sufficient to promote a carcinogenic process. This conjecture should be explored and tested - mathematically and statistically, with the availability of genomic data on databanks. In this work, we explore the model proposed by Bozic and collaborators (2010) that describes the evolution of the tumour according to a Galton-Watson process. Besides, the model gives the relation between the numbers of passenger mutations giving a specific number of driver mutations. We intend to explore some of the model parameters and test some premises about the number of drive mutations and selective advantage, comparing the simulation results with genomic data from colorectal cancer patients. The genomic data was obtained from the DBMutation (http://www.bioinformatics-brazil.org/dbmutation/), a comprehensive database for genomic mutations in cancer. We expect that correlations between driver mutations and the time evolution of tumour process will facilitate the interpretation of genomic information, to make them useful and applicable to clinical oncology.