PALS db is a collection of Putative Alternative Splicing information from 19 936 human UniGene clusters and 16 615 mouse UniGene clusters. Alternative splicing (AS) sites were predicted by using the longest messenger RNA (mRNA) sequence in each UniGene cluster as the reference sequence. This sequence was aligned with related sequences in UniGene and dbEST to reveal the AS. This information was presented with six features: (i) literature aliases were used to improve the result of a gene name search; (ii) the quality of a prediction can be easily judged from the color-coded similarity and the scaled length of an alignment; (iii) we have clustered those EST sequences that support the same AS site together to enhance the users’ confidence on a prediction; (iv) the users can also set up the alignment criteria interactively to recover false negatives; (v) tissue distribution can be displayed by placing the mouse cursor over an alignment; (vi) gene features will be analyzed at foreign sites by submitting the selected mRNA or its encoded protein as a query. Using these features, the users cannot only discover putative AS sites in silico, but also make new observations by combining AS information with tissue distributions or with gene features. PALS db is available at http://palsdb.ym.edu.tw/.

Background: Genome-wide disease-gene finding approaches may sometimes provide us with a long list of candidate genes. Since using pure experimental approaches to verify all candidates could be expensive, a number of network-based methods have been developed to prioritize candidates. Such tools usually have a set of parameters pre-trained using available network data. This means that re-training network-based tools may be required when existing biological networks are updated or when networks from different sources are to be tried. Results: We developed a parameter-free method, interconnectedness (ICN), to rank candidate genes by assessing the closeness of them to known disease genes in a network. ICN was tested using 1,993 known disease-gene associations and achieved a success rate of ~44% using a protein-protein interaction network under a test scenario of simulated linkage analysis. This performance is comparable with those of other well-known methods and ICN outperforms other methods when a candidate disease gene is not directly linked to known disease genes in a network. Interestingly, we show that a combined scoring strategy could enable ICN to achieve an even better performance (~50%) than other methods used alone. Conclusions: ICN, a user-friendly method, can well complement other network-based methods in the context of prioritizing candidate disease genes.

Background: Recombinant protein production is a useful biotechnology to produce a large quantity of highly soluble proteins. Currently, the most widely used production system is to fuse a target protein into different vectors in Escherichia coli (E. coli). However, the production efficacy of different vectors varies for different target proteins. Trial-and-error is still the common practice to find out the efficacy of a vector for a given target protein. Previous studies are limited in that they assumed that proteins would be over-expressed and focused only on the solubility of expressed proteins. In fact, many pairings of vectors and proteins result in no expression. Results: In this study, we applied machine learning to train prediction models to predict whether a pairing of vector-protein will express or not express in E. coli. For expressed cases, the models further predict whether the expressed proteins would be soluble. We collected a set of real cases from the clients of our recombinant protein production core facility, where six different vectors were designed and studied. This set of cases is used in both training and evaluation of our models. We evaluate three different models based on the support vector machines (SVM) and their ensembles. Unlike many previous works, these models consider the sequence of the target protein as well as the sequence of the whole fusion vector as the features. We show that a model that classifies a case into one of the three classes (no expression, inclusion body and soluble) outperforms a model that considers the nested structure of the three classes, while a model that can take advantage of the hierarchical structure of the three classes performs slight worse but comparably to the best model. Meanwhile, compared to previous works, we show that the prediction accuracy of our best method still performs the best. Lastly, we briefly present two methods to use the trained model in the design of the recombinant protein production systems to improve the chance of high soluble protein production. Conclusion: In this paper, we show that a machine learning approach to the prediction of the efficacy of a vector for a target protein in a recombinant protein production system is promising and may compliment traditional knowledge-driven study of the efficacy. We will release our program to share with other labs in the public domain when this paper is published.

Ganoderma lucidum ( G. lucidum ) is a medicinal mushroom renowned in East Asia for its potential biological effects. To enable a systematic exploration of the genes associated with the various phenotypes of the fungus, the genome consortium of G. lucidum has carried out an expressed sequence tag (EST) sequencing project. Using a Sanger sequencing based approach, 47,285 ESTs were obtained from in vitro cultures of G. lucidum mycelium of various durations. These ESTs were further clustered and merged into 7,774 non-redundant expressed loci. The features of these expressed contigs were explored in terms of over-representation, alternative splicing, and natural antisense transcripts. Our results provide an invaluable information resource for exploring the G. lucidum transcriptome and its regulation. Many cases of the genes over-represented in fast-growing dikaryotic mycelium are closely related to growth, such as cell wall and bioactive compound synthesis. In addition, the EST-genome alignments containing putative cassette exons and retained introns were manually curated and then used to make inferences about the predominating splice-site recognition mechanism of G. lucidum . Moreover, a number of putative antisense transcripts have been pinpointed, from which we noticed that two cases are likely to reveal hitherto undiscovered biological pathways. To allow users to access the data and the initial analysis of the results of this project, a dedicated web site has been created at http://csb2.ym.edu.tw/est/ .

Background: MicroRNAs are very small non-coding RNAs that interact with microRNA recognition elements (MREs) on their target messenger RNAs. Varying the concentration of a given microRNA may influence the expression of many target proteins. Yet, the expression of a specific target protein can be fine-tuned by alternative cleavage and polyadenylation to the corresponding mRNA.Results: This study showed that alternative splicing of mRNA is a fine-tuning mechanism in the cellular regulatory network. The splicing-regulated MREs are often highly repressive MREs. This phenomenon was observed not only in the hsa-miR-148a-regulated DNMT3B gene, but also in many target genes regulated by hsa-miR-124, hsa-miR-1, and hsa-miR-181a. When a gene contains multiple MREs in transcripts, such as the VEGF gene, the splicing-regulated MREs are again the highly repressive MREs. Approximately one-third of the analysable human MREs in MiRTarBase and TarBase can potentially perform the splicing-regulated fine-tuning. Interestingly, the high (+30%) repression ratios observed in most of these splicing-regulated MREs indicate associations with functions. For example, the MRE-free transcripts of many oncogenes, such as N-RAS and others may escape microRNA-mediated suppression in cancer tissues.Conclusions: This fine-tuning mechanism revealed associations with highly repressive MRE. Since high-repression MREs are involved in many important biological phenomena, the described association implies that splicing-regulated MREs are functional. A possible application of this observed association is in distinguishing functionally relevant MREs from predicted MREs.

Background: In biological systems, pathways coordinate or interact with one another to achieve a complex biological process. Studying how they influence each other is essential for understanding the intricacies of a biological system. However, current methods rely on statistical tests to determine pathway relations, and may lose numerous biologically significant relations.Results: This study proposes a method that identifies the pathway relations by measuring the functional relations between pathways based on the Gene Ontology (GO) annotations. This approach identified 4,661 pathway relations among 166 pathways from Pathway Interaction Database (PID). Using 143 pathway interactions from PID as testing data, the function-based approach (FBA) is able to identify 93% of pathway interactions, better than the existing methods based on the shared components and protein-protein interactions. Many well-known pathway cross-talks are only identified by FBA. In addition, the false positive rate of FBA is significantly lower than others via pathway co-expression analysis.Conclusions: This function-based approach appears to be more sensitive and able to infer more biologically significant and explainable pathway relations.