The general theme of my research is to develop computational frameworks that recognize biologically meaningful patterns implicated in high-throughput (HT) array and next-generation sequencing (NGS) data. The following motivations pertains to the more specific research perspectives of mine in transcriptomics, comparative epigenomics, and characterization of cancer genomes. The “omics” are only relevant in the context of HT data analysis. However, the corresponding computational challenges in fact appear as a recurrent theme in many other (unrelated) research areas. Currently the "building blocks" for NGS analyses are still under active development. Unlike microarray, the technical variability of NGS is largely unknown. Sequencing bias, read distribution, background sequences, and many more present great challenges to bioinformaticians. Meanwhile, it's an exciting arena that has lots of room for contributions by “borrowing” ideas from fundamental works established in statistics and computer science.


For instance, numerous machine learning and statistical literatures have described the analysis of sequential data, in which data points are not independent identically distributed (i.i.d). These data points either continuous (e.g. in speech recognition) or discrete (e.g. prediction of next outbreak of pandemic disease) may represent distinct generative processes (e.g. true signal v.s. background noise). Probabilistic frameworks such as Hidden Markov model (HMM) is a popular tool to infer the underlying states that best explain the data. More advanced Bayesian approaches taking into account the (non-homogenous) stochastic processes are less prone to overfitting than the maximum likelihood frameworks. However, the latter application is limited by their computational overheads and (usually) requires approximation approaches such as Markov Chain Monte Carlo (aka sampling), variational inference (with some simplifying assumptions), or empirical Bayes (empirical estimation of the priors from marginal posterior) to perform efficient learning and inference. Adaptations of these frameworks in the context of NGS analyses are at the core of my research.


Inferring microRNA regulatory networks

MicroRNAs (miRNAs) are ~22 nucleotide long noncoding RNA species. The regulatory roles of microRNAs (miRNA) have important implication in developments and diseases. Functional characterization of miRNAs require accurate identifications of their RNA targets, which has been a challenging computational task due to various confounding factors centering around the combinatorial co-regulatory relationships between miRNA and mRNA. Earlier developed sequence-based methods are mostly based on seed match, phylogenetic conservation, and binding energy. Recently, there is a paradigm shift from the sequence-based binary classification to more quantitative expression-based and network -focused approach. The momentum of this shift is largely facilitated by the increasing amount of expression profiling data of mRNAs and miRNAs across various experimental conditions. My current main research is to infer cancer-specific miRNA regulatory networks that  can characterize cancer phenotypes and/or facilitate prognostic biomarkers development.


  1. •Li, Y., Liang, C., Wong, KC., Luo, J., Zhang Z. (May, 2014). Mirsynergy: detecting synergistic miRNA regulatory modules by overlapping neighbourhood expansion. Bioinformatics

  2. •Li, Y., Liang, C., Wong, KC, Jin, K., and Zhang, Z. (Feb, 2014) Inferring probabilistic miRNA-mRNA interaction signatures in cancers: a role-switch approach. Nucleic Acids Research, 42(9), e76. doi: 10.1093/nar/gku182

  3. •Li, Y., Goldenberg, A., Wong, KC., Zhang Z. (Oct, 2013). A probabilistic approach to explore human miRNA targetome by integrating miRNA-overexpression data and sequence information. Bioinformatics (Oxford, England), 30(5), 621–628. doi:10.1093/bioinformatics/btt599


RNA epigenetics

N6-methyladenosine (m6A) is the most prevalent endogenous methylation in RNA. Recently, Dominissini et al. (2010) and Mayer et al. (2010) have demonstrated a novel NGS protocol to interrogate transcriptome-wide m6A methylation using m6A-seq, based on antibody-mediated capture and massively parallel sequencing. Despite implicated in regulation of gene expression, the functional roles of m6A are still largely unknown. In collaboration with Prof. Crystal Zhao, we are exploring deeper the fundamental biology of m6A in mammalian development with combined experimental and computational approach.


  1. •Wang, Y., Li, Y., Toth, J. I., Petroski, M. D., Zhang, Z., & Zhao, J. C. (2014). N6-methyladenosine modification destabilizes developmental regulators in embryonic stem cells. Nature Cell Biology, 16(2), 1-10. doi:10.1038/ncb2902


Detection of protein-associated noncoding RNA from RIP-seq, CLIP-seq, and PAR-CLIP experiments

Comprehensive transcriptome analyses suggest that only 1%-2% of the human or mouse genome is protein coding whereas 70%-90% is transcriptionally active, but do not code for proteins, and thus denoted as non-coding RNA (ncRNA) (ENCODE Project Consortium, 2007). Mounting evidence suggests that many of these ncRNAs are evolutionarily conserved, functionally interact with transcription factors and/or chromatin regulators, and participate in gene regulation. NGS platforms such as PAR-CLIP and RIP-Seq enables unbiased genome-wide identification of these ncRNAs and thus promise to reveal unique aspects of molecular biology. We are closely working with biologists from CCBR to construct protein-protein and protein-ncRNA interaction networks utilizing these technologies.


  1. •Zhao, D., Li, Y., Greenblatt, J., & Zhang, Z. (2014). ncRNA–Protein Interactions in Development and Disease from the Perspective of High-Throughput Studies. In A. Emili, J. Greenblatt, & S. Wodak (Eds.), Systems Analysis of Chromatin-Related Protein Complexes in Cancer (pp. 87-115). Springer New York. doi:10.1007/978-1-4614-7931-4_5

  2. •Li, Y., Zhao, D. Y., Greenblatt, J. F., & Zhang, Z. (2013). RIPSeeker: a statistical package for identifying protein-associated transcripts from RIP-seq experiments. Nucleic Acids Research, 41(8), e94. doi:10.1093/nar/gkt142


Identification of differential DNA methylation and copy number variation in cancer

In DNA methylation, a hydrogen atom of the cytosine base of the DNA is replaced by a methyl group. This change typically induces a locally more compact chromatin structure, repressing gene activities in the vicinity. On the other hand, copy number variation (CNV) correspond to deletion or duplication of large regions of the genome relative to normal subjects. Abnormal DNA methylation pattern and CNV in cancer have been reported in many studies. In a collaborative project with Prof. Art Petronis from The Krembil Family Epigenetics Laboratory, we use tiling arrays to interrogate both aforementioned abnormal phenomena in sera from large cohorts of colorectal cancer patients to establish some prominent (epi-)genetic signatures. In the future, we will be using (bisulfite) NGS to confirm and extend our current findings.


Past works

We proposed and implemented a computational pipeline to analyze peptide array kinome data (Li et al., 2012). The work as my B.Sc. Honours thesis was under supervision of Dr. Anthony Kusalik and in collaboration with immunologists (co-authors) from the Vaccine and Infectious Disease Organization (VIDO) at the U of S.


To our knowledge, the proposed pipeline is the first integrative approach that addresses kinome-specific computational challenges in microarray analyses. In particular, our statistical testing for differentially phosphorylated kinase peptides takes into account the technical and biological variation inherent to the technology and dynamic kinase activities between biological replicates, respectively. Comparing to existing methods, our approach is more sensitive in detecting kinases involved in well-defined signaling pathways activated by the select stimuli. The central roles of kinases in immune defence make them promising therapeutic targets. Rigorous detection of subtle changes in treatment-specific kinase activities via a powerful platform such as kinome microarray may facilitate pharmaceutical design against diseases.

  1. •Arsenault, R. J., Li, Y., Maattanen, P., Scruten, E., Doig, K., Potter, A., Griebel, P., Kusalik, A., and Napper, S. (2013) Altered Toll-like receptor 9 signaling in Mycobacterium avium subsp. paratuberculosis infected bovine monocytes reveals potential therapeutic targets. Infection and immunity, 81(1), 226-237.

  2. •Arsenault, R. J., Li, Y., Potter, A., Griebel, P. J., Kusalik, A., and Napper, S. (2012). Induction of ligand-specific PrPC signaling in human neuronal cells. Prion, 6(5), 477-488.

  3. •Arsenault, R. J., Li, Y., Bell, K., Doig, K., Potter, A., Griebel, P. J., Kusalik, A., and Napper, S. (2012). Mycobacterium avium subsp. paratuberculosis Inhibits Interferon Gamma-Induced Signaling in Bovine Monocytes. Insights into the Cellular Mechanisms of Johne’s Disease. Infection and immunity, 80, 3039–3048.

  4. •Li, Y., Arsenault, R. J., Trost, B., Slind, J., Griebel, P. J., Napper, S., and Kusalik, A. (2012). A Systematic Approach for Analysis of Peptide Array Kinome Data. Science Signaling, 5(220), pl2–pl2.