A5034394920- RNA modifications and cancer
- RNA Research and Splicing
- RNA and protein synthesis mechanisms
- Cancer-related molecular mechanisms research
- Neural Networks and Applications
- Parallel Computing and Optimization Techniques
- Nuclear and radioactivity studies
University of Toronto
2014-2018
Ontario Genomics
2018
Canadian Institute for Advanced Research
2014
To facilitate precision medicine and whole-genome annotation, we developed a machine-learning technique that scores how strongly genetic variants affect RNA splicing, whose alteration contributes to many diseases. Analysis of more than 650,000 intronic exonic revealed widespread patterns mutation-driven aberrant splicing. Intronic disease mutations are 30 nucleotides from any splice site alter splicing nine times as often common variants, missense have the least impact on protein function...
Alternative splice site selection is inherently competitive and the probability of a given to be used also depends on strength neighboring sites. Here, we present new model named (COSSMO), which explicitly accounts for these effects predicts percent selected index (PSI) distribution over any number putative We an alternative splicing event as choice 3' acceptor conditional fixed upstream 5' donor or site. build four different architectures that use convolutional layers, communication long...
Abstract When estimating expression of a transcript or part using RNA-seq data, it is commonly assumed that reads are generated uniformly from positions within the transcript. While this assumption acceptable for long sequences where many averaged, frequently leads to large errors short sequences, e.g ., less than 100 bp. Analysis such as when studying splice junctions and microRNAs, increasingly important necessitates addressing in short-sequence estimation. Indeed, we examined data diverse...
Abstract Motivation Alternative splice site selection is inherently competitive and the probability of a given to be used also depends strongly on strength neighboring sites. Here we present new model named Competitive Splice Site Model (COSSMO), which explicitly models these effects predict PSI distribution over any number putative We an alternative splicing event as choice 3’ acceptor conditional fixed upstream 5’ donor site, or site. build four different architectures that use...