Oncogenes expand during evolution to withstand somatic amplification.

Background: Cancer-related genes are under intense evolutionary pressure. We conjectured that gene size is an important determinant of amplification propensity for oncogenes and thus cancer susceptibility and therefore could be subject to natural selection. Patients and methods: Gene information, including size and genomic locations, of all protein-coding genes were downloaded from Ensembl (release 87). Quantification of gene amplification was based on Genomic Identification of Significant Targets in Cancer scores obtained from available The Cancer Genome Atlas studies. Results: Oncogenes are larger in size as compared with non-cancer genes (mean size: 92.1 kb versus 61.4 kb; P < 0.0001) in the human genome, which is contributed by both increased total exon size (mean size: 4.6 kb versus 3.4 kb; P < 0.0001) and higher intronic content (mean %: 84.8 versus 78.0; P < 0.01). Such non-random size distribution and intronic composition are conserved in mouse and Drosophila (all P < 0.0001). Stratification by gene age indicated that young oncogenes have been subject to a stronger evolutionary pressure for gene expansion than their non-cancer counterparts. Pan-cancer analysis demonstrated that larger oncogenes were amplified to a lesser extent. Tumor-suppressor genes also moved toward small oncogenes in the course of evolution. Conclusions: Oncogenes expand in size whereas tumor-suppressor genes move closer to small oncogenes in the course of evolution to withstand oncogenic somatic amplification. Our findings have shed new light on the previously unappreciated influence of gene size on oncogene amplification and elucidated how cancers have shaped our genome to its present configuration.

Detecting reliable non interacting proteins (NIPs) significantly enhancing the computational prediction of protein-protein interactions using machine learning methods.

Protein-protein interactions (PPIs) play a vital role in most biological processes. Hence their comprehension can promote a better understanding of the mechanisms underlying living systems. However, besides the cost and the time limitation involved in the detection of experimentally validated PPIs, the noise in the data is still an important issue to overcome. In the last decade several in silico PPI prediction methods using both structural and genomic information were developed for this purpose. Here we introduce a unique validation approach aimed to collect reliable non interacting proteins (NIPs). Thereafter the most relevant protein/protein-pair related features were selected. Finally, the prepared dataset was used for PPI classification, leveraging the prediction capabilities of well-established machine learning methods. Our best classification procedure displayed specificity and sensitivity values of 96.33% and 98.02%, respectively, surpassing the prediction capabilities of other methods, including those trained on gold standard datasets. We showed that the PPI/NIP predictive performances can be considerably improved by focusing on data preparation.

Diffusion on curved, periodic surfaces.

We present a simulation algorithm for a diffusion on a curved surface given by the equation phi(r)=0. The algorithm is tested against analytical results known for diffusion on a cylinder and a sphere, and applied to the diffusion on the P, D, and G periodic nodal surfaces. It should find application in an interpretation of two-dimensional exchange NMR spectroscopy data of diffusion on biological membranes.