Uneven size distribution of mammalian genes in the number of tissues expressed and in the number of co-expressed genes

Hope for GWAS: relevant risk genes uncovered from GWAS statistical noise

Hundreds of genetic variants have been associated to common diseases through genome-wide association studies (GWAS), yet there are limits to current approaches in detecting true small effect risk variants against a background of false positive findings. Here we addressed the missing heritability problem, aiming to test whether there are indeed risk variants within GWAS statistical noise and to develop a systematic strategy to retrieve these hidden variants. Employing an integrative approach, which combines protein-protein interactions with association data from GWAS for 6 common diseases, we found that associated-genes at less stringent significance levels (p < 0.1) with any of these diseases are functionally connected beyond noise expectation. This functional coherence was used to identify disease-relevant subnetworks, which were shown to be enriched in known genes, outperforming the selection of top GWAS genes. As a proof of principle, we applied this approach to breast cancer, supporting well-known breast cancer genes, while pinpointing novel susceptibility genes for experimental validation. This study reinforces the idea that GWAS are under-analyzed and that missing heritability is rather hidden. It extends the use of protein networks to reveal this missing heritability, thus leveraging the large investment in GWAS that produced so far little tangible gain.

In plants, expression breadth and expression level distinctly and non-linearly correlate with gene structure

BACKGROUND

Compactness of highly/broadly expressed genes in human has been explained as selection for efficiency, regional mutation biases or genomic design. However, highly expressed genes in flowering plants were shown to be less compact than lowly expressed ones. On the other hand, opposite facts have also been documented that pollen-expressed Arabidopsis genes tend to contain shorter introns and highly expressed moss genes are compact. This issue is important because it provides a chance to compare the selectionism and the neutralism views about genome evolution. Furthermore, this issue also helps to understand the fates of introns, from the angle of gene expression.

RESULTS

In this study, I used expression data covering more tissues and employ new analytical methods to reexamine the correlations between gene expression and gene structure for two flowering plants, Arabidopsis thaliana and Oryza sativa. It is shown that, different aspects of expression pattern correlate with different parts of gene sequences in distinct ways. In detail, expression level is significantly negatively correlated with gene size, especially the size of non-coding regions, whereas expression breadth correlates with non-coding structural parameters positively and with coding region parameters negatively. Furthermore, the relationships between expression level and structural parameters seem to be non-linear, with the extremes of structural parameters possibly scale as power-laws or logrithmic functions of expression levels.

CONCLUSION

In plants, highly expressed genes are compact, especially in the non-coding regions. Broadly expressed genes tend to contain longer non-coding sequences, which may be necessary for complex regulations. In combination with previous studies about other plants and about animals, some common scenarios about the correlation between gene expression and gene structure begin to emerge. Based on the functional relationships between extreme values of structural characteristics and expression level, an effort was made to evaluate the relative effectiveness of the energy-cost hypothesis and the time-cost hypothesis.

Alternative polyadenylation variants of the RNA binding protein, HuR: abundance, role of AU-rich elements and auto-Regulation

The RNA-binding protein, HuR, is involved in the stabilization of AU-rich element-containing mRNAs with products that are involved in cell-cycle progression, cell differentiation and inflammation. We show that there are multiple polyadenylation variants of HuR mRNA that differ in their abundance, using both bioinformatics and experimental approaches. A polyadenylation variant with distal poly(A) signal is a rare transcript that harbors functional AU-rich elements (ARE) in the 3'UTR. A minimal 60-nt region, but not a mutant form, fused to reporter-3'UTR constructs was able to downregulate the reporter activity. The most predominant and alternatively polyadenylated mature transcript does not contain the ARE. HuR itself binds HuR mRNA, and upregulated the activity of reporter from constructs fused with ARE-isoform and the HuR ARE. Wild-type tristetraprolin (TTP), but not the zinc finger mutant TTP, competes for HuR binding and upregulation of HuR mRNA. The study shows that the HuR gene codes for several polyadenylation variants differentially regulated by AU-rich elements, and demonstrates an auto-regulatory role of HuR.