Olga Botvinnik

Variation in homeodomain DNA-binding revealed by high-resolution analysis of sequence preferences

Most homeodomains are unique within a genome, yet many are highly conserved across vast evolutionary distances, implying strong selection on their precise DNA-binding specificities. We determined the binding preferences of the majority (168) of mouse homeodomains to all possible 8-base sequences, revealing rich and complex patterns of sequence specificity and showing that there are at least 65 distinct homeodomain DNA-binding activities. We developed a computational system that successfully predicts binding sites for homeodomain proteins as distant from mouse as Drosophila and C. elegans, and we infer full 8-mer binding profiles for the majority of known animal homeodomains. Our results provide an unprecedented level of resolution in the analysis of this simple domain structure and suggest that variation in sequence recognition may be a factor in its functional diversity and evolutionary success.

Characterizing genomic alterations in cancer by complementary functional associations

Systematic efforts to sequence the cancer genome have identified large numbers of mutations and copy number alterations in human cancers. However, elucidating the functional consequences of these variants, and their interactions to drive or maintain oncogenic states, remains a challenge in cancer research. We developed REVEALER, a computational method that identifies combinations of mutually exclusive genomic alterations correlated with functional phenotypes, such as the activation or gene dependency of oncogenic pathways or sensitivity to a drug treatment. We used REVEALER to uncover complementary genomic alterations associated with the transcriptional activation of β-catenin and NRF2, MEK-inhibitor sensitivity, and KRAS dependency. REVEALER successfully identified both known and new associations, demonstrating the power of combining functional profiles with extensive characterization of genomic alterations in cancer genomes.

Dysregulation of RBFOX2 is an early event in cardiac pathogenesis of diabetes

Alternative splicing (AS) defects that adversely affect gene expression and function have been identified in diabetic hearts; however, the mechanisms responsible are largely unknown. Here, we show that the RNA-binding protein RBFOX2 contributes to transcriptome changes under diabetic conditions. RBFOX2 controls AS of genes with important roles in heart function relevant to diabetic cardiomyopathy. RBFOX2 protein levels are elevated in diabetic hearts despite low RBFOX2 AS activity. A dominant-negative (DN) isoform of RBFOX2 that blocks RBFOX2-mediated AS is generated in diabetic hearts. DN RBFOX2 interacts with wild-type (WT) RBFOX2, and ectopic expression of DN RBFOX2 inhibits AS of RBFOX2 targets. Notably, DN RBFOX2 expression is specific to diabetes and occurs at early stages before cardiomyopathy symptoms appear. Importantly, DN RBFOX2 expression impairs intracellular calcium release in cardiomyocytes. Our results demonstrate that RBFOX2 dysregulation by DN RBFOX2 is an early pathogenic event in diabetic hearts.

Prediction of response to therapy with ezatiostat in lower risk myelodysplastic syndrome

BackgroundApproximately 70% of all patients with myelodysplastic syndrome (MDS) present with lower-risk disease. Some of these patients will initially respond to treatment with growth factors to improve anemia but will eventually cease to respond, while others will be resistant to growth factor therapy. Eventually, all lower-risk MDS patients require multiple transfusions and long-term therapy. While some patients may respond briefly to hypomethylating agents or lenalidomide, the majority will not, and new therapeutic options are needed for these lower-risk patients. Our previous clinical trials with ezatiostat (ezatiostat hydrochloride, Telentra®, TLK199), a glutathione S-transferase P1-1 inhibitor in clinical development for the treatment of low- to intermediate-risk MDS, have shown significant clinical activity, including multilineage responses as well as durable red-blood-cell transfusion independence. It would be of significant clinical benefit to be able to identify patients most likely to respond to ezatiostat before therapy is initiated. We have previously shown that by using gene expression profiling and grouping by response, it is possible to construct a predictive score that indicates the likelihood that patients without deletion 5q will respond to lenalidomide. The success of that study was based in part on the fact that the profile for response was linked to the biology of the disease.MethodsRNA was available on 30 patients enrolled in the trial and analyzed for gene expression on the Illumina HT12v4 whole genome array according to the manufacturer’s protocol. Gene marker analysis was performed. The selection of genes associated with the responders (R) vs. non-responders (NR) phenotype was obtained using a normalized and rescaled mutual information score (NMI).ConclusionsWe have shown that an ezatiostat response profile contains two miRNAs that regulate expression of genes known to be implicated in MDS disease pathology. Remarkably, pathway analysis of the response profile revealed that the genes comprising the jun-N-terminal kinase/c-Jun molecular pathway, which is known to be activated by ezatiostat, are under-expressed in patients who respond and over-expressed in patients who were non-responders to the drug, suggesting that both the biology of the disease and the molecular mechanism of action of the drug are positively correlated.

GOATOOLS: A Python library for Gene Ontology analyses

The biological interpretation of gene lists with interesting shared properties, such as up- or down-regulation in a particular experiment, is typically accomplished using gene ontology enrichment analysis tools. Given a list of genes, a gene ontology (GO) enrichment analysis may return hundreds of statistically significant GO results in a “flat” list, which can be challenging to summarize. It can also be difficult to keep pace with rapidly expanding biological knowledge, which often results in daily changes to any of the over 47,000 gene ontologies that describe biological knowledge. GOATOOLS, a Python-based library, makes it more efficient to stay current with the latest ontologies and annotations, perform gene ontology enrichment analyses to determine over- and under-represented terms, and organize results for greater clarity and easier interpretation using a novel GOATOOLS GO grouping method. We performed functional analyses on both stochastic simulation data and real data from a published RNA-seq study to compare the enrichment results from GOATOOLS to two other popular tools: DAVID and GOstats. GOATOOLS is freely available through GitHub: https://github.com/tanghaibao/goatools.