Westmead Hub Bioinformatics and Omics Seminars

The seminar aims to engage researchers within the field of bioinformatics, data science, and omics. The seminars are held at 10:00 am on the last Thursday of each month and broadcast using Zoom. The format of the seminar is 45 minutes presentation and 10 minutes of questions.

2023 Seminar Series – Now held online

Contact hani.kim@sydney.edu.au for more information and to be added to the mailing list.

2022 Seminars

31st March, 2022

Speaker: David Croucher

Title: Memory of stochastic single-cell apoptotic signalling promotes chemoresistance in neuroblastoma

Abstract

David completed his undergraduate studies and PhD research at the University of Wollongong, followed by a post-doctoral position at the Garvan Institute, focusing on the functional characterisation of individual proteins in the behaviour of cancer cells. Pursuing the idea that these cancer-related proteins do not act in isolation, but are instead embedded within dynamic networks, David undertook a second post-doctoral position at Systems Biology Ireland in University College Dublin.

After establishing a research group at Systems Biology Ireland, David received a Future Research Leader grant from Cancer Institute NSW and returned to the Garvan Institute as head of the Network Biology group. His research group now focuses on combining computational modelling with novel proteomic and single-cell technologies to investigate how alterations in dynamic signalling networks can lead to tumour progression and therapeutic resistance.

The full recording can be found here: https://youtu.be/l7oZvl4QNLY

Past Seminars

2021 Seminars

Thursday, October 28, 2021

Speaker: Irina Voineagu

Title: Comprehensive evaluation of deconvolution methods for human brain gene expression

Abstract

Gene expression measurements, similar to DNA methylation and proteomic measurements, are influenced by the cellular composition of the sample analysed. Deconvolution of bulk transcriptome data aims to estimate the cellular composition of a sample from its gene expression data, which in turn can be used to correct for composition differences across samples. Although a multitude of deconvolution methods have been developed, it is unclear whether their performance is consistent across tissues with different complexities of cellular composition. The human brain is unique in its transcriptomic diversity, expressing the highest diversity of alternative splicing isoforms and non-coding RNAs. It comprises a complex mixture of cell-types including transcriptionally similar sub-types of neurons, which undergo gene expression changes in response to neuronal activity. However, a comprehensive assessment of the accuracy of transcriptome deconvolution methods on human brain data is currently lacking.

We carried out the first comprehensive comparative evaluation of the accuracy of deconvolution methods for human brain transcriptome data, and assess the tissue-specificity of our key observations by comparison with transcriptome data from human pancreas and heart.

We evaluate 8 transcriptome deconvolution approaches, covering all main classes: 4 partial deconvolution methods, each applied with 9 different cell-type signatures, 2 enrichment methods, and 2 complete deconvolution methods. We test the accuracy of cell-type estimates using in silico mixtures of single-cell RNA-seq data, mixtures of neuronal and glial RNA, as well as nearly 2,000 human brain samples.

Our results bring several important insights into the performance of transcriptome deconvolution: (a) We find that cell-type signature data has a stronger impact on brain deconvolution accuracy than the choice of method. (b) We demonstrate that biological factors influencing brain cell-type signature data (e.g. brain region, in vitro cell culturing), have stronger effects on the deconvolution outcome than technical factors (e.g. RNA sequencing platform). (c) We find that partial deconvolution methods outperform complete deconvolution methods on human brain data. To facilitate wider implementation of correction for cellular composition, we develop a webtool that implements the best performing methods, and is available at https://voineagulab.shinyapps.io/BrainDeconvShiny/ .

Thursday, September 30, 2021

Alan F Rubin, Bioinformatics Division, Walter and Eliza Hall institute

Title: Understanding the functional consequences of every single mutation in Your Favourite Gene

Abstract: A central problem in genomic medicine is understanding the effects of individual DNA variants. Although their function can be explored in the lab, assessing variants individually is both time and resource intensive. Multiplexed Assays of Variant Effect (MAVEs) are a family of experimental techniques that allow researchers to measure all possible variants of a gene or other functional element at once. This field has expanded dramatically in recent years and is poised to become an integral part of clinical variant interpretation. Since a wide community of scientists and clinicians use these data, the results must be reproducible, auditable and widely available. This talk will focus on the computational biology and bioinformatics of MAVEs with an emphasis on deep mutational scanning (DMS), which targets protein-coding genes. I will cover the way DMS datasets are analysed and current efforts to build on our field-leading methods, as well as new machine learning applications for DMS score prediction and imputation. I will also describe MaveDB, the database of record for MAVEs, and upcoming improvements to the sharing of MAVE data for clinical use.

Thursday, June 24, 2021

Warwick Locke, CSIRO

Title: Beyond Cancer: Circulating DNA in injury and illness

Abstract: Circulating cell free DNA (ccfDNA) has long been used in the oncology domain, where it is used to study the underlying genotype of a tumour without the need for invasive procedures (e.g. circulating tumour DNA). However, cell free DNA molecules also contain a wealth of information outside cancer. Early studies show that ccfDNA quantitates in the plasma vary with injury and/or illnesses and that ccfDNA concentrations strongly correlate with severity and clinical outcome. Cell free DNA levels in a healthy individual are typically low and produced predominately by the natural turnover of blood cell types. In injury and illness, trauma and inflammation result in pronounced cell death and the release of abnormal amounts of ccfDNA from the affected tissue(s). The use of ccfDNA in injury is made more difficult by the lack of easy to detect genetic mutations as found in Cancer. Therefore, another layer of information is required.

The DNA methylome is highly specific to cell type, varying across the various tissues of an organism despite the identical underlying genetic sequence. We are developing DNA methylation-based biomarkers for the detection of tissue specific cell death using PCR. Here I will discuss the approaches we take in identifying candidate loci and optimising them for clinical translation using PCR based methods.

Thursday, May 27, 2021

Seyhan Yazar, Garvan Institute of Medical Research (Garvan-Weizmann Centre for Cellular Genomics)

Title: Single-cell eQTL mapping identifies cell type specific genetic control of autoimmune disease

Abstract: Dr Seyhan Yazar is an early career researcher and a computational biologist. She received her PhD from the University of Western Australia in 2016. During her PhD studies, she worked on a range of research projects involving design, methods, theory and analysis around genetic and environmental influences in common complex diseases and associated traits, through exploring data from population-based studies. Upon completion of her PhD studies, she received an NHMRC CJ Martin Biomedical Fellowship. As part of her her fellowship, Dr Yazar undertook specialist training in bioinformatics under the supervision of Professor Colin Semple at the MRC Human Genetics Unit of the Institute of Genetics and Molecular Medicine at the University of Edinburgh. Applying core bioinformatic principles to emerging sequencing technologies, she completed the publicly available genome assembly of the bare-nosed wombat. Returning to Australia, she joined Associate Professor Joseph Powell’s laboratory at the Garvan-Weizmann Centre for Cellular Genomics. In second half of her fellowship, she has completed the largest single cell expression quantitative trait loci (eQTL) study of immune cells identifying cell type specific genetic control of autoimmune disease. Her current focus is multidimensional single cell analysis to identify biomarkers of immune-mediated complex diseases.

Thursday, April 29, 2021

Leila Eshraghi, Garvan Institute of Medical Research (Connie Johnson Breast Cancer Research)

Title: Genomic Autopsy of Perinatal Death

Abstract: In Australia, six babies are born still every day, and despite recent advances in both scientific and medical expertise, the rate of stillbirth has not declined in two decades. The Genomic Autopsy Study (GAS) is a National research team based in Adelaide, South Australia that has been undertaking genomic research to try and understand the genetic contribution to stillbirth and perinatal death. Accurate diagnosis of the cause of PD is essential for appropriate counselling, including risk to future pregnancies in families, and is essential for the provision of reproductive options such as pre-implantation genetic diagnosis (PGD). Our study team not only aims to investigate how; changes to the fetal (and parental) DNA, presence of fetal infection (either maternal transmission across the placenta, or fetal infection via the birth canal), and/or changes to the placenta, can contribute to stillbirth and perinatal death but also aims to perform a change-in-management study to determine the health and economic benefits of future PD avoided births with and genomic testing as SOC. To date, 134 families have been engaged in our Genomic Autopsy Study (through their referring clinical service), and our data indicates that we can increase diagnosis and prevent recurrence in at least 50% of cases, and thus could make a significant contribution towards the target of reducing the rate of stillbirth.

Thursday, February 25, 2021

Laurence Wilson, CSIRO (Transformational Bioinformatics)

Title: Genomics and Biosecurity: Developing new methods for defending against emerging threats

Abstract: Genomics is playing an increasingly important role in the fight against biosecurity threats including invasive species, anti-microbial resistance and emerging pathogens. The ability to curate, query and track the spread of genotypes amongst a wild population is critical for enabling targeted intervention strategies. However the constantly changing genetic landscapes as well as the wealth of data being generated present several key challenges. Among these include identifying unknown changes in poorly annotated genomes and handling the wealth of genomic data being generated. To address this, our team has developed several strategies including alignment free approaches for identifying when foreign DNA has integrated into a host’s genome and cloud-based platforms for easy and secure sharing of genomic data. With a broad range of approaches, our team aims to provide researchers, clinicians and policy makers with the tools they need to make targeted interventions to combat emerging threats.