Exploring the functional impact of alternative splicing on human protein isoforms using available annotation sources

Sulakhe, Dinanath; D’Souza, Mark; Wang, Sheng; Balasubramanian, Sandhya; Athri, Prashanth; Xie, Bingqing; Canzar, Stefan; Agam, Gady; Gilliam, T. Conrad; Maltsev, Natalia

doi:10.1093/bib/bby047

Cited by 24 publications

(22 citation statements)

References 137 publications

(157 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A recent review (Sulakhe et al , 2018) reported a set of function-related sequence features (as defined by UniProt; Breuza et al , 2016) associated with a list of isoforms. The presence or absence of these functional sequence features can be used to infer potential isoform functions.…”

Section: Results and Validationmentioning

confidence: 99%

“…Isoforms carry specific, sometimes distinct or even opposing, biological functions. Moreover, the expression of an isoform is often specific to tissue, developmental stage or environmental conditions, which is responsible for the diversity and adaptability of cellular activities (Sulakhe et al , 2018; Wang et al , 2008). Therefore, delineating the functions of isoforms is crucial to the study of functional complexity and diversity of genomes.…”

Section: Introductionmentioning

confidence: 99%

“…Different types of biological data may carry complementary information of isoform functions, and hence a systematic integration of such information might lead to a substantial improvement in prediction accuracy (Li et al , 2016; Sulakhe et al , 2018). In particular, we may divide informative biological data into the following two types.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

DIFFUSE: predicting isoform functions from sequences and expression profiles via deep learning

et al. 2019

View full text Add to dashboard Cite

Motivation Alternative splicing generates multiple isoforms from a single gene, greatly increasing the functional diversity of a genome. Although gene functions have been well studied, little is known about the specific functions of isoforms, making accurate prediction of isoform functions highly desirable. However, the existing approaches to predicting isoform functions are far from satisfactory due to at least two reasons: (i) unlike genes, isoform-level functional annotations are scarce. (ii) The information of isoform functions is concealed in various types of data including isoform sequences, co-expression relationship among isoforms, etc. Results In this study, we present a novel approach, DIFFUSE (Deep learning-based prediction of IsoForm FUnctions from Sequences and Expression), to predict isoform functions. To integrate various types of data, our approach adopts a hybrid framework by first using a deep neural network (DNN) to predict the functions of isoforms from their genomic sequences and then refining the prediction using a conditional random field (CRF) based on co-expression relationship. To overcome the lack of isoform-level ground truth labels, we further propose an iterative semi-supervised learning algorithm to train both the DNN and CRF together. Our extensive computational experiments demonstrate that DIFFUSE could effectively predict the functions of isoforms and genes. It achieves an average area under the receiver operating characteristics curve of 0.840 and area under the precision–recall curve of 0.581 over 4184 GO functional categories, which are significantly higher than the state-of-the-art methods. We further validate the prediction results by analyzing the correlation between functional similarity, sequence similarity, expression similarity and structural similarity, as well as the consistency between the predicted functions and some well-studied functional features of isoform sequences. Availability and implementation https://github.com/haochenucr/DIFFUSE. Supplementary information Supplementary data are available at Bioinformatics online.

show abstract

Section: Results and Validationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

DIFFUSE: predicting isoform functions from sequences and expression profiles via deep learning

et al. 2019

View full text Add to dashboard Cite

show abstract

“…It is ensured that both proteins of valid PPIs that constitute search hits in PROTEINATOR are co-expressed in at least one common tissue, and members in a common biological pathway. Large-scale multi-omics projects have identified coordinated expression patterns of proteins based on tissue type (a list and summary in Sulakhe et al((Sulakhe et al, 2018))). PROTEINATOR uses these contextual co-localizations of interacting proteins as a condition to identify candidate PPIs.…”

Section: Introductionmentioning

confidence: 99%

PROTEINATOR: Web-UI exploring repurposing hypotheses of PROTEIN InhibiTORs based on protein interactions

Tangadu

Shankar

Varanasi

et al. 2019

Preprint

Self Cite

View full text Add to dashboard Cite

The authors have withdrawn this manuscript because in it's current stage, the work only identifies the plausible search space for new drug interactions, and is based on a biochemical phenomenon. Nonetheless, and as pointed out by 2 independent peer review processes from leading journals in this area, the results are of very little value without basic validation. We are a computer science department, and do not have the experimental setup to do the same.As we move forward, our goal is to collaborate with a life sciences (biophysical lab) group, and apply for funding to complete the validations for a few test cases before we re-submit this work and make it available to the rest of the community. Therefore, the authors do not wish this work to be cited as reference for the project. If you have any questions, please contact the corresponding author.

show abstract

“…In parallel, molecular studies have been conducted to understand the mechanisms behind the dynamic changes in event patterns, identifying a large number of RNA binding proteins as regulators of AltTP [9][10][11][12][13][14] . In response to the recognition of the biological importance of AltTP, bioinformatics tools have been developed to analyze the structural and regulatory aspects of AltTP events and have contributed to the description and understanding of AltTP (reviewed in 15 ).…”

mentioning

confidence: 99%

tappAS: a comprehensive computational framework for the analysis of the functional impact of differential splicing

Fuente

Arzalluz-Luque

Tardáguila

et al. 2019

Preprint

View full text Add to dashboard Cite

Traditionally, the functional analysis of gene expression data has used pathway and network enrichment algorithms. These methods are usually gene rather than transcript centric and hence fall short to unravel functional roles associated to posttranscriptional regulatory mechanisms such as Alternative Splicing (AS) and Alternative PolyAdenylation (APA), jointly referred here as Alternative Transcript Processing (AltTP). Moreover, short-read RNA-seq has serious limitations to resolve full-length transcripts, further complicating the study of isoform expression. Recent advances in long-read sequencing open exciting opportunities for studying isoform biology and function. However, there are no established bioinformatics methods for the functional analysis of isoform-resolved transcriptomics data to fully leverage these technological advances. Here we present a novel framework for Functional Iso-Transcriptomics analysis (FIT). This framework uses a rich isoform-level annotation database of functional domains, motifs and sites -both coding and noncoding-and introduces novel analysis methods to interrogate different aspects of the functional relevance of isoform complexity. The Functional Diversity Analysis (FDA) evaluates the variability at the inclusion/exclusion of functional domains across annotated transcripts of the same gene. Parameters can be set to evaluate if AltTP partially or fully disrupts functional elements. FDA is a measure of the potential of a multiple isoform transcriptome to have a functional impact. By combining these functional labels with expression data, the Differential Analysis Module evaluates the relative contribution of transcriptional (i.e. gene level) and post-transcriptional (i.e. transcript/protein levels) regulation on the biology of the system. Measures of inclusion of NLS, transmembrane domains or DNA binding motifs, for example.Some of these findings were experimentally validated by others and us.In summary, we propose a novel framework for the functional analysis of transcriptomes at isoform resolution. We anticipate the tappAS tool will be an important resource for the adoption of the Functional Iso-Transcriptomics analysis by functional genomics community.

show abstract

Exploring the functional impact of alternative splicing on human protein isoforms using available annotation sources

Cited by 24 publications

References 137 publications

DIFFUSE: predicting isoform functions from sequences and expression profiles via deep learning

DIFFUSE: predicting isoform functions from sequences and expression profiles via deep learning

PROTEINATOR: Web-UI exploring repurposing hypotheses of PROTEIN InhibiTORs based on protein interactions

tappAS: a comprehensive computational framework for the analysis of the functional impact of differential splicing

Contact Info

Product

Resources

About