RNA-seq Based Transcription Characterization of Fusion Breakpoints as a Potential Estimator for Its Oncogenic Potential.

Based on high-throughput sequencing technology, the detection of gene fusions is no longer a big challenge but estimating the oncogenic potential of fusion genes remains challenging. Recent studies successfully applied machine learning methods and gene structural and functional features of fusion mutation to predict their oncogenic potentials. However, the transcription characterizations features of fusion genes have not yet been studied. In this study, based on the clonal evolution theory, we hypothesized that a fusion gene is more likely to be an oncogenic genomic alteration, if the neoplastic cells harboring this fusion mutation have larger clonal size than other neoplastic cells in a tumor. We proposed a novel method, called iFCR (internal Fusion Clone Ratio), given an estimation of oncogenic potential for fusion mutations. We have evaluated the iFCR method in three public cancer transcriptome sequencing datasets; the results demonstrated that the fusion mutations occurring in tumor samples have higher internal fusion clone ratio than normal samples. And the most frequent prostate cancer fusion mutation, TMPRSS2-ERG, appears to have a remarkably higher iFCR value in all three independent patients. The preliminary results suggest that the internal fusion clone ratio might potentially advantage current fusion mutation oncogenic potential prediction methods.

Multiclass classification of sarcomas using pathway based feature selection method.

Feature selection is an important research topic in bioinformatics, to date a large number of methods have been developed. Recently several pathway based feature selection protocols, such as the condition-responsive genes method, have been proposed for better classification performance. However, these conventional pathway based methods may lead to the selection of relevant but redundant genes in a given pathway while missing the other useful genes. Also these methods were limited to binary classification, while in many clinical problems a multiclass protocol is preferred such as the classification of sarcomas. Here, we propose a new pathway based feature selection method named Redundancy Removable Pathway based feature selection method (RRP) for the binary and multiclass classification problems. Three classifiers were implemented to compare the performance and gene functions of gene-based, conventional pathway based, and our RRP method. The validation results suggest that the RRP method is a feasible and robust feature selection method for multi-class prediction problems.

Hox genes from the parasitic flatworm Schistosoma japonicum.

Hox genes are characterized by a highly conserved peptide domain and contribute to antero-posterior axis patterning during embryogenesis. These genes have been widely studied in a variety of animal species due to their central role in evolutionary developmental biology. Based on the published genome assembly and unpublished re-sequencing project data, we present the first genome-wide characterization and comparative genomic analysis of the Hox gene family within Schistosoma japonicum. Eight Hox genes were identified and validated in our investigation. Phylogenetic analysis revealed that these genes are distributed among seven orthology groups of the Hox gene family. Our study further suggested that differences in the Lox5 gene copy number existed between the two closely related species, S. japonicum and Schistosoma mansoni. Semi-quantitative real-time polymerase chain reaction experiments revealed that Lox5 and Hox4 gene expression was high in the schistosomulum stage, and all four genes investigated showed highest expression within the eggs.

Analysis of floral transcription factors from Lycoris longituba.

Transcription factors (TFs) are proteins that bind to specific promoter regions of their target genes and regulate gene transcription. Many of these factors have been found to influence flowering. Lycoris longituba exhibits a great deal of diversity in flower color and flower form, making it a suitable model for the study of floral development. We have identified 338 putative TFs from more than thirty thousand ESTs sequenced from the floral tissue of L. longituba, and validated them using real-time RT-PCR. Fifty-one of the TFs were recognized as being potentially flower-specific, and the expression patterns of some of them during six flowering phases have been elucidated. Homolog annotation and phylogenetic analysis revealed that some TFs that belong to several TF families, such as MADS, MYB-related, NAC, and ABI3-VP1, were suggested to play important roles in the flowering process. Our dataset may be used to identify priority target TF genes for further study.