BEGIN:VCALENDAR VERSION:2.0 PRODID:researchseminars.org CALSCALE:GREGORIAN X-WR-CALNAME:researchseminars.org BEGIN:VEVENT SUMMARY:Peter Kharchenko (Harvard University) DTSTART:20210503T141000Z DTEND:20210503T144000Z DTSTAMP:20260422T212827Z UID:ComputationalBiology/1 DESCRIPTION:Title: Bayesian segmentation of spatially resolved transcriptomics d ata\nby Peter Kharchenko (Harvard University) as part of Computational Biology Symposium\n\n\nAbstract\nSpatial transcriptomics is an emerging s tack of technologies\, which adds spatial dimension to conventional single -cell RNA-sequencing. New protocols\, based on in situ sequencing or multi plexed RNA fluorescent in situ hybridization register positions of single molecules in fixed tissue slices. Analysis of such data at the level of in dividual cells\, however\, requires accurate identification of cell bounda ries. While many existing methods are able to approximate cell center posi tions using nuclei stains\, current protocols do not report robust signal on the cell membranes\, making accurate cell segmentation a key barrier fo r downstream analysis and interpretation of the data. To address this chal lenge\, we developed a tool for Bayesian Segmentation of Spatial Transcrip tomics Data (Baysor)\, which optimizes segmentation considering the likeli hood of transcriptional composition\, size and shape of the cell. The Baye sian approach can take into account nuclear or cytoplasm staining\, howeve r can also perform segmentation based on the detected transcripts alone. W e show that Baysor segmentation can in some cases nearly double the number of the identified cells\, while reducing contamination. Importantly\, we demonstrate that Baysor performs well on data acquired using five differen t spatially-resolved protocols\, making it a useful general tool for analy sis of high-resolution spatial data.\n LOCATION:https://researchseminars.org/talk/ComputationalBiology/1/ END:VEVENT BEGIN:VEVENT SUMMARY:Smita Krishnaswamy (Yale University) DTSTART:20210503T144500Z DTEND:20210503T151500Z DTSTAMP:20260422T212827Z UID:ComputationalBiology/2 DESCRIPTION:Title: Geometric and Topological Approaches to Representation Learni ng in Biomedical Data\nby Smita Krishnaswamy (Yale University) as part of Computational Biology Symposium\n\n\nAbstract\nHigh-throughput\, high- dimensional data has become ubiquitous in the biomedical\, health and soci al sciences as a result of breakthroughs in measurement technologies and d ata collection. While these large datasets containing millions of observat ions of cells\, peoples\, or brain voxels hold great potential for unders tanding generative state space of the data\, as well as drivers of differe ntiation\, disease and progression\, they also pose new challenges in term s of noise\, missing data\, measurement artifacts\, and the so-called “c urse of dimensionality.” In this talk\, I will cover data geometric and topological approaches to understanding the shape and structure of the dat a. First\, we show how diffusion geometry and deep learning can be used to obtain useful representations of the data that enable denoising (MAGIC) \, dimensionality reduction (PHATE)\, and factor analysis (Archetypal Anal ysis Network) of the data. Next we will show how to learn dynamics from s tatic snapshot data by using a manifold-regularized neural ODE-based optim al transport (TrajectoryNet). Finally\, we cover a novel approach to combi ne diffusion geometry with topology to extract multi-granular features fro m the data (Diffusion Condensation and Multiscale PHATE) to assist in diff erential and predictive analysis. On the flip side\, we also create a mani fold geometry from topological descriptors\, and show its applications to neuroscience. Together\, we will show a complete framework for exploratory and unsupervised analysis of big biomedical data.\n LOCATION:https://researchseminars.org/talk/ComputationalBiology/2/ END:VEVENT BEGIN:VEVENT SUMMARY:Meromit Singer (Harvard Medical School) DTSTART:20210503T152000Z DTEND:20210503T155000Z DTSTAMP:20260422T212827Z UID:ComputationalBiology/3 DESCRIPTION:Title: Utilizing coupled single-cell RNA-seq and TCR-seq to reveal T h17 systemic dynamics during homeostasis and disease\nby Meromit Singe r (Harvard Medical School) as part of Computational Biology Symposium\n\n\ nAbstract\nIn this talk we will describe a systemic study of the transcrip tional and clonal characteristics and function of Th17 cells throughout mu ltiple mouse organs\, as revealed by coupled single-cell RNA-seq and TCR-s eq\, and validated in follow-up experiments in the lab. We will describe h ow we utilized 84\,000 tissue Th17 cells profiled during homeostasis and d isease to characterize their heterogeneity\, plasticity\, and migration at homeostasis and during CNS autoimmunity. We discovered a homeostatic Th17 cell population\, that is induced by the intestinal microbiota\, is prese nt in both lymphoid organs and the intestine\, and expresses IL-17. We dis covered that during EAE this homeostatic population gives rise to a pathog enic Th17 cell population\, that migrates specifically through the drainin g lymph nodes and the spleen to the CNS\, and highly expresses a specific subset of cytokines. \nIn this talk we will emphasize how coupled single-c ell RNA-seq and TCR data was used to generate hypotheses regarding cell su btype characterization and T cell clonality and migration\, and how such h ypotheses were followed-up on experimentally.\n LOCATION:https://researchseminars.org/talk/ComputationalBiology/3/ END:VEVENT BEGIN:VEVENT SUMMARY:John Marioni (EMBL-EBI) DTSTART:20210503T155500Z DTEND:20210503T162500Z DTSTAMP:20260422T212827Z UID:ComputationalBiology/4 DESCRIPTION:by John Marioni (EMBL-EBI) as part of Computational Biology Sy mposium\n\nAbstract: TBA\n LOCATION:https://researchseminars.org/talk/ComputationalBiology/4/ END:VEVENT BEGIN:VEVENT SUMMARY:Uri Alon (Weizmann Institute) DTSTART:20210503T172500Z DTEND:20210503T175500Z DTSTAMP:20260422T212827Z UID:ComputationalBiology/5 DESCRIPTION:Title: Design principles of hormone circuits\nby Uri Alon (Weizm ann Institute) as part of Computational Biology Symposium\n\nAbstract: TBA \n LOCATION:https://researchseminars.org/talk/ComputationalBiology/5/ END:VEVENT BEGIN:VEVENT SUMMARY:Eran Segal (Weizmann Institute) DTSTART:20210503T180000Z DTEND:20210503T183000Z DTSTAMP:20260422T212827Z UID:ComputationalBiology/6 DESCRIPTION:Title: Harnessing big data for personalized medicine\nby Eran Se gal (Weizmann Institute) as part of Computational Biology Symposium\n\n\nA bstract\nThe recent availability of diverse health data resources on large cohorts of human individuals presents many challenges and opportunities. I will present our work aimed at developing machine learning algorithms fo r predicting future onset of disease and identifying causal drivers of dis ease based on nationwide electronic health record data as well as data fro m high-throughput omics profiling technologies such as genetics\, microbio me\, and metabolomics. Our models provide novel insights into potential dr ivers of obesity\, diabetes\, and heart disease\, and identify hundreds of novel markers at the microbiome\, metabolite\, and immune system level. O verall\, our predictive models can be translated into personalized disease prevention and treatment plans\, and to the development of new therapeuti c modalities based on metabolites and the microbiome.\n LOCATION:https://researchseminars.org/talk/ComputationalBiology/6/ END:VEVENT BEGIN:VEVENT SUMMARY:Martin Hemberg (Brigham and Women’s Hospital) DTSTART:20210503T183500Z DTEND:20210503T190500Z DTSTAMP:20260422T212827Z UID:ComputationalBiology/7 DESCRIPTION:Title: Searching for alien DNA – characterization of sequences tha t are not present in the DNA\nby Martin Hemberg (Brigham and Women’s Hospital) as part of Computational Biology Symposium\n\n\nAbstract\nNullo mers and nullpeptides are short DNA or amino acid sequences that are absen t from a genome or proteome\, respectively. One potential cause for their absence could be that they have a detrimental impact on an organism. Here\ , we identified all possible nullomers and nullpeptides in the genomes and proteomes of over thirty species and show that a significant proportion o f these sequences are under negative selection. We assign nullomers to dif ferent functional categories (coding sequences\, exons\, introns\, 5’UTR \, 3’UTR\, regulatory regions and promoters) and show that nullomers fro m coding sequences and promoters are most likely to be selected against. S imilarly\, we find that regulatory regions and transcription factor bindin g sites harbor more mutations resulting in nullomers than expected. Furthe r analysis of coding regions also reveals specific pathways where mutation s are more likely to result in nullomers or nullpeptides. Utilizing varian ts in the human population\, we annotate variant-associated nullomers\, hi ghlighting their potential use as DNA ‘fingerprints’.\n LOCATION:https://researchseminars.org/talk/ComputationalBiology/7/ END:VEVENT BEGIN:VEVENT SUMMARY:Elana Fertig (Johns Hopkins) DTSTART:20210503T191000Z DTEND:20210503T194000Z DTSTAMP:20260422T212827Z UID:ComputationalBiology/8 DESCRIPTION:Title: Uncovering hidden sources of transcriptional dysregulation ar ising from inter- and intra-tumor heterogeneity.\nby Elana Fertig (Joh ns Hopkins) as part of Computational Biology Symposium\n\n\nAbstract\nHete rogeneity poses a major challenge in translational research. For example\, inter-tumor heterogeneity limits the biomarker discovery and intra-tumor heterogeneity enables therapeutic resistance. Moreover\, in some cancers d river mutations are insufficient to account for the widespread transcripti onal variation responsible for these outcomes. Thus\, new computational to ols to model transcriptional variation are essential. To address this we d evelop an innovative computational framework\, Expression Variation Analys is (EVA)\, to model transcriptional dysregulation in cancer. Briefly\, EVA quantifies transcriptional heterogeneity for one set of samples or cells from one phenotype using the expected dissimilarity between pairs of expre ssion profiles. U-statistics theory can then quantify the statistical sign ificance of the difference in transcriptional heterogeneity between phenot ypes. We apply EVA to perform a comprehensive characterization of transcri ptional variation in head and neck squamous cell carcinoma (HNSCC). At a p athway level\, transcriptional variation in HNSCC tumors is higher than no rmal controls. Applying EVA to integrate ChIP-seq data with RNA-seq reveal s that these pervasive transcriptional differences occur in enhancers. Sim ilarly\, applying EVA at a gene level to model splicing reveals more heter ogeneity in transcript usage in tumor samples than normals. HPV- HNSCC tum ors are unique in having mutations in genes that regulate the splicing mac hinery\, and the HPV- tumors with these alterations have a greater number of dysregulated splice variants than those without. Nonetheless\, the EVA analysis identifies a similar number of alternative splice variants in HPV + as HPV- tumors suggesting an alternative mechanism of transcriptional he terogeneity in HPV+ disease. Adapting EVA to single cell data demonstrates that increased fibroblast composition is associated with greater variatio n in immune pathway activity in HNSCC. Moreover\, we observe greater trans criptional heterogeneity in HNSCC primary tumors than lymph node metastasi s consistent with a clonal outgrowth. We demonstrate that the statistical framework from EVA enables differential heterogeneity analysis in HNSCC ra nging from pathway dysregulation\, splice variation\, epigenetic regulatio n\, and single cell analysis. This algorithm provides a critical framework to model the hidden multi-molecular mechanisms underlying the complex pat ient outcomes that are pervasive in cancer.\n LOCATION:https://researchseminars.org/talk/ComputationalBiology/8/ END:VEVENT END:VCALENDAR