ResearchMatch 4.0 - initial projects submissions
Please find below the list of ResearchMatch projects received from the McGill life science and clinical communities. The McGill data science community is now being asked to contact the individual researchers if interested to collaborate.
For additional information about each project proposal such as the Project description, Description of datasets to be used or generated, Why you see this as a collaborative research project and what you hope to gain from the collaboration, log in to the ResearchMatch portal.
Projects are split into 3 groups:
Clinical Research (8) Life Science Research (16) Population Health Research (3)
Clinical Research
Project title: Deep mining electronic health record data to guide COVID-19 care
We propose to apply machine learning methods to anonymized data extracted from the MUHC data warehouse for patients admitted for COVID-19. The primary focus is to use temporal ML methods to exploit historical data and understand how prior medical conditions influence the clinical trajectory of these patients, including their response to therapy.
Keywords: machine learning, datawarehouse, clinical prediction, COVID-19
PI: David Buckeridge
Department: Epidemiology and Biostatistics
email: david.buckeridge [at] mcgill.ca
Project title: EV biomarkers in saliva and blood from head & neck cancer patients by ChIP-seq & RNA-seq
Background: Head and neck squamous cell carcinoma (HNSCC) is the sixth most common cancer worldwide in 2018. HNSCC can be distinguished into HPV-negative and HPV-positive HNSCC. Limited biomarkers in clinical settings and careful physical examination remains the primary approach for detection. Therefore, new biomarkers are needed for early diagnosis and rapid treatment. Exosomes are emerging as biomarkers and many studies have shown the miRNA-based biomarkers in the plasma/serum of head and neck cancer (HNC).
Keywords: liquid biopsy, extracellular vesicles, head and neck cancer, exosomes, ChIP-seq, RNA-seq, biomarker
PI: Julia Burnier
Department: Oncology and Pathology
email: Julia.burnier [at] mcgill.ca
Project title: Development of a Severity Scoring System for Intermediate/Mild Zellweger Spectrum Disorders
Zellweger Spectrum Disorders (ZSD) are a group of rare, autosomal recessive, multisystemic disorders characterized by a defect in peroxisome biogenesis. ZSD is a phenotypic continuum ranging from severe to mild, however, severity of symptoms cannot be entirely predicted by genotype or biochemical abnormalities. The lack of a robust tool to assess disease severity is a major limitation in the management and treatment this disorder. Our objective is to develop and validate a robust and quantitative severity scoring system based on the body systems that can be affected in individuals with intermediate/mild ZSD. A validated ZSD severity scoring system would allow clinicians to identify prognostic features in ZSD patients and reliably quantify disease burden. Moreover, it could be used for classification and stratification of patients in clinical trials that seek to evaluate potential treatments for ZSD, to distinguish clinical endpoints, and for monitoring disease progression and treatment responses.
Keywords: Zellweger Spectrum Disorder, severity scoring system, peroxisome biogenesis disorders, rare diseases, disease management, prognosis, clinical trial stratification
PI: Nancy Braverman
Department: Department of Paediatrics
email: pbd.genetics [at] mcgill.ca
Project title: Impact of autism spectrum disorder on functional connectivity in intracranial recordings
Autism spectrum disorder (ASD) has been shown to be associated with aberrant functional connectivity. Although evidence strongly support long-range underconnectivity, the effect of ASD on short-range functional connectivity is unclear, mostly due to the low spatial resolution of non-invasive recording modalities like EEG. To better understand how ASD impacts neuronal interactions, we will pool intracranial recordings from ASD patients that underwent surgery for drug-resistant epilepsy. We will use these recordings to compute the phase locking index – a measure of functional connectivity – between pairs of local and distal electrodes. We expect to show that inter-hemispheric connectivity is reduced and local functional connectivity is increased in ASD compared to non-ASD patients.
Keywords: neural dynamics, epilepsy, autism, neurodevelopmental disorders, intracranial recording, functional connectivity
PI: Mayada Elsabbagh
Department: Neurology and Neurosurgery
email: mayada.elsabbagh [at] mcgill.ca
Project title: A Novel Approach to Improve the Identification of Unstable Plaques and the Prediction of Strokes
Strokes, caused by the rupture of unstable atherosclerotic plaques in the carotid arteries, are leading causes of death world-wide. Unfortunately, the current method used in clinical practice to predict whether a plaque will rupture is limiting and based solely on the degree of artery stenosis. This has led to suboptimal medical decisions and inappropriate treatment allocation. Recent research has shown that the plaque’s composition is a better predictor of clinical outcomes than artery stenosis. Herein, we aim to improve the identification of unstable plaques and the prediction of strokes through the development of risk scores by combining ultrasonic plaque features, patient clinical information, and blood markers, in addition to artery stenosis.
Keywords: artificial intelligence, Image analysis and machine learning, ultrasound, stroke, atherosclerotic plaque, risk score
PI: Stella Daskalopoulou
Department: Medicine
email: stella.daskalopoulou [at] mcgill.ca
Project title: Withdrawal of renin-angiotensin blockade in patients with advanced kidney disease: the USRDS dataset
Medications that block the renin-angiotensin-aldosterone system have been shown to reduce all cause-mortality and a composite outcome of cardiovascular death, non-fatal myocardial infarction or stroke among patients with heart failure and/or after a myocardial infarction. They have also been shown to slow down progression of chronic kidney disease. However, use of these agents is controversial in patients with significantly impaired kidney function. Although several studies attempted to estimate their potential benefits (or risks) in this population, they had important limitations. Therefore, information is lacking or contradictory with respect to their cessation or continuation among patients with advanced kidney disease. This is a critical question because of the very high morbidity or mortality associated with kidney disease and the increasing prevalence of this condition in the general population. Effective and safe treatment strategies are urgently needed for these patients.
In this study, we will use the United States Renal Data System database, the largest of its kind in the world, to assess whether continuation or withdrawal of medication that block the renin-angiotensin-aldosterone system is associated with better clinical outcomes among patients with advanced kidney disease.
Keywords: Big data analysis, myocardial infarction , kidney health, Heart Failure
PI: Thomas Mavrakanas
Department: Medicine
email: thomas.mavrakanas [at] mcgill.ca
Project title: Quantification and Reproducibility of Cerebellar Atrophy Patterns in an Essential Tremor Cohort
Essential tremor (ET) is one of the most common chronic neurological movement disorders with a prevalence of 5% in people over age 60. It presents as a postural or action tremor that may affect one of both upper limbs. The tremor may be disabling and in many cases has a familial component. Medical treatment may be effective, and in some cases surgical treatment targeting the thalamus can be offered as an effective alternative. Although the underlying pathophysiology of ET remains unknown, recent postmortem studies have revealed Purkinje cell loss in the cerebellar gray matter and the surrounding neuronal cells. Neuroimaging studies have provided variable results, with some work pointing to volume loss in cerebellar grey matter, and others revealing mainly normal morphological features on MRI scans. Furthermore, it remains unclear whether specific tremor patterns and severity correlate with local neural loss or macro-atrophy patterns.
In this project, we aim to rigorously assess cerebellar atrophy patterns using MR imaging data from a local anonymized clinical dataset comprising 38 well-characterized ET patients and 32 normal controls (NC). We will perform multi-scale structural analyses to quantify anatomical changes at voxel (VBM), surface (cortical thickness), and regional (cerebellar lobules) levels. We will determine whether regional volume loss or altered voxel intensity correlates in different cerebellar regions correlates with tremor severity. We will also compare these findings with local and public MR-imaging cohorts of another movement disorder, Parkinson’s disease, that can also present with tremor, and be difficult to distinguish from ET especially at early stages. Subsequently, we will perform several methodological vibration analyses to assess the reproducibility of our results and possibly to explain variability pertaining to cerebellar involvement in ET that exists in the current literature. Last, we will explore how to best use larger open datasets to gain statistical power for the analysis of clinical data with small sample size.
Keywords: Essential Tremor, Parkinson‚ Disease, Neuroimaging, Reproducible Neuroscience, Brain Imaging Analyses, Statistical Analyses, Confounder Correction, Cerebellum
PI: Abbas Sadikot
Department: Neurology and Neurosurgery
email: abbas.sadikot [at] mcgill.ca
Project title: Deep Learning to Predict Disease Progression in Patients with Polycystic Kidney Disease
Autosomal dominant polycystic kidney disease (ADPKD) is the most common inherited kidney disease, and the fourth leading cause of kidney failure leading to dialysis or kidney transplantation. There exists significant phenotypic heterogeneity, both within and between families, even among those with the same gene mutation. As a result, clinical risk prediction has focused on estimating total kidney volume (TKV), which reflects the overall burden of cyst development and expansion. Those with higher TKV at a younger age are more likely to be considered for disease modifying therapy. The estimation of TKV by radiologic imaging is practical, but volumetric measurement alone does not sufficiently classify risk, and may lead to inappropriate decisions related to treatment. Recognizing certain patterns of kidney cyst involvement is critical, but this pattern recognition is not standardized and not easily measurable. Deep learning approaches to image analysis of kidneys in patient with ADPKD may provide a more reliable way to assess patients at risk of developing renal progression and subsequently identifying those most amenable to targeted therapy.
Keywords: imaging, deep learning, Image analysis and machine learning, risk prediction, kidney health
PI: Ahsan Alam
Department: Medicine
email: ahsan.alam [at] mcgill.ca
Life Science Research
Project title: Genetic Interaction Networks in Breast Cancer
Breast cancer is the most common cancer in women worldwide and the second leading cause of death by cancer. In 2020, 27,400 women were diagnosed with breast cancer, representing 25% of all new cancer diagnosis. Triple-negative (TN) breast cancer (lacking expression of the markers for the estrogen, progesterone, and human epidermal growth factor receptor 2 (HER2) receptors) represent ~15% of these cases. TNBC is the most aggressive in term of tumour growth and metastasis and is associated with earlier age of onset and the worse overall outcome. This is in part due to the complexity of alterations that ocucur in TNBC and still there is poor understanding of those which are causal. This project develops an approach to examine complex mutational events at scale
Keywords: None
PI: Morag Park
Department: Oncology, Biochemistry, Goodman Cancer Centre
email: morag.park [at] mcgill.ca
Project title: Analysis of 2D and 3D models of neural electrical activity in Parkinson's disease models
Parkinson's disease (PD), a central nervous system disorder that leads to dopaminergic (DA) neuronal cell death, has no disease-modifying treatments. The pathways leading to DA neuronal cell death and PD patient phenotypes are not understood. Some genes leading to familial forms of PD are known. Mouse models with deletions or mutations of genes causing familial PD do not model the disease phenotypes seen in humans. The ideal way to study the cellular mechanisms leading to human disease is to use human cells. Taking advantage of several technological breakthroughs, the Fon lab uses human blood cells, reprogrammed into induced pluripotent stem cells (hiPSCs). Real brain tissue contains a mixture of cell types, and to better model brain tissue we use 3D models termed midbrain organoids (hMOs), which contain DA neurons and a mixture of other brain cells. We have recorded functional activity present in hMOs using Multi-electrode-array (MEA) recordings from familial models of PD and controls. The midbrain DA neurons in the human substania nigra, the part of the brain affected by PD, fire in a consistent bursting pattern and we detect a similar pattern in the hMOs. We aim to distinguish phenotypic differences between neurons and hMOs derived from monogenic models of PD and the matching isogenic controls. To do this we must develop scripts and algorithms to measure features from the raw MEA data. Additionally, very little is known about the functional properties of iPSC-derived DA neurons or hMOs. We aim to develop analysis systems to quantify signal and network dynamics in these systems.
Keywords: neural circuits, Parkinson Disease, electrophysiology, neural organoids, Multi-Electrode Array (MEA), neural electrical activity, network dynamics
PI: Edward Fon
Department: Neurology & Neurosurgery
email: ted.fon [at] mcgill.ca
Project title: Analysis of DNA variation introduced by reprogramming blood cells into pluripotent stem cells
Human induced pluripotent stem cells (hiPSCs) made from human somatic cells have opened up new avenues to model human diseases and for developing novel therapeutics. Using hiPSCs-derived cells for drug discovery requires extensive quality control measures and characterization. Growing hiPSCs for a long time, followed by DNA sequencing, has revealed that changes in the sequence of housekeeping genes and other genetic changes, like duplication and deletion events occur over time. Changes in DNA sequences during reprogramming from somatic cells into hiPSCs have also been reported. However, these changes have not been well investigated. We performed whole genome sequencing on pairs of somatic and reprogrammed cells. We aim to quantify the number of new genetic variants seen after reprogramming and determine if there are any trends in these changes across 19 patient samples.
Keywords: genomics, Rare variants, Parkinson Disease, induced pluripotent stem cells, iPSC, whole genome sequencing, ALS
PI: Thomas Durcan
Department: Neurology and Neurosurgery
email: thomas.durcan [at] mcgill.ca
Project title: Data and tools for deciphering the neuronal basis of learning and recollection strategies in mice
Modern innovations in neural recording have revolutionized our ability to decipher the neural dynamics that govern cognition and behaviour in a variety of animal models. Specifically, advances in calcium imaging enable tracking the activity of hundreds of cells across timescales spanning multiple days, and miniaturized microscopes have brought this technique to the forefront of behavioural neuroscience to study the development of neural dynamics in freely behaving animals. Such ambitious tools accelerate large-scale data collection, but with large datasets comes the challenge of efficiently extracting signals from noise and structures from constituents. Deep learning approaches in data science have optimized the handling of large datasets through a variety of neural network systems and computational models specifically designed to tackle high-dimensional data. It is thus through this intersection between modern neurophysiologists in the Williams laboratory and computational neuroscientists from the Richards laboratory that we hope to collaborate on a new project aiming to decipher the neuronal basis of learning and recollection strategies in mice. A great challenge in mapping the neural representation of learning in the brain stems from the fact that memory does not function as a localized, unitary entity. More accurately, conscious declarative memory relies on hippocampal-mediated systems, whereas unconscious habitual memory relies on striatal-mediated systems [1]. While specific tests exist to study either system in isolation, both systems interact in practice to optimize learning [2]. Understanding the relative contributions of both systems is ideal for mapping neurophysiology to behaviour. Using recently collected data in mice on a behavioural test designed to isolate spatial memory processes in rodents, we propose a collaborative project to study how different learning strategies predict the alternating usage and combined interactions of multiple memory systems in the brain. We are hopeful that any insights gained from the data and analytical tools created can be readily shared with the greater neuroscience community through our established research platform at the Douglas Research Centre [3].
[1] Squire, L. R., J. Cognit. Neurosci. 4, 3 (1992). [2] Poldrack, R. A., et al., Neuropsycholgia. 41, 3 (2003). [3] Mosser et al., Genes Brain Behav. 20, 1 (2021).
Keywords: deep learning, neural circuits, calcium imaging, memory, freely moving mice
PI: Blake Richards
Department: Neurology & Neurosurgery/Computer Science
email: blake.richards [at] mcgill.ca
Project title: Predictability of responses in the human hippocampus to context-free static images
Hippocampus is often studied in reference to its critical role in episodic memory and spatial navigation. Recent findings highlighting the role of the hippocampus in perceptual tasks such as scene perception and attention have challenged the traditional view of its function as being purely memory-related. Despite this, not much is known about how the hippocampus responds to visual stimuli like those that are often used to study the visual and scene processing cortices. Here we aim to study the hippocampus responses to randomly displayed static images devoid of temporal and spatial contexts and assess the similarity of the response patterns in this area to other precursor areas in the visual cortex and parahippocampal cortex as well as to computational models of object recognition.
Keywords: memory, Vision, hippocampus , fmri
PI: Pouya bashivan
Department: Physiology
email: pouya.bashivan [at] mcgill.ca
Project title: Molecular mechanisms during allergic lung disease
The type 2 immune response is critical for host defense against large parasites such as helminths, wound healing and body metabolism. On the other hand, dysregulation of type 2 immunity causes immunopathological conditions, including asthma, atopic dermatitis, rhinitis, and anaphylaxis, which have been referred to as type 2 immunopathologies. Thus, a balanced type 2 immune response must be achieved to mount effective protection against invading pathogens while avoiding immunopathology. The proposed research aims to decipher the mechanisms between innate and adaptive immune regulation at the single cell level and will as such give detailed insights into the molecular regulation of type 2 immunopathologies.
Keywords: Allergic lung disease, Asthma, group 2 innate lymphoid cells (ILC2), CD4 T cells, RNASeq, single-cell RNA sequencing
PI: Jorg Fritz
Department: Microbiology & Immunology
email: jorg.fritz [at] mcgill.ca
Project title: Evaluating the role of nuclear genes encoding mitochondrial and tRNA proteins in neurodegeneration
Systemic mitochondrial dysfunction has been described in neurodegenerative diseases and may arise as a consequence of abnormal mitochondrial DNA or due to abnormal interacting proteins. Specifically, mutations in genes directly implicated in mitochondrial function such as Parkin, PINK1, and SOD1 cause rare Parkinson’s disease (PD) and amyotrophic lateral sclerosis (ALS), respectively. Furthermore, accumulation of another PD protein, alpha-synuclein, has been linked to mitochondrial impairment. Finally, mitochondrial dysfunction has also been described in atypical forms of PD such as multiple system atrophy. It remains unclear whether mitochondrial dysfunction initiates or is a byproduct of, neurodegeneration. Therefore, conducting a large-scale unbiased study to evaluate whether there is enrichment of variation in nuclearly encoded mitochondrial proteins as well as in tRNA enzymes in neurodegenerative disease cases, can help resolve the role of mitochondria in neurodegeneration.
Keywords: disease cohorts, genetics, statistical analyses, Population genetics, Human genetics, neuro
PI: Sali Farhan
Department: Neurosurgery and Neurology, and Human Genetics
email: sali.farhan [at] mcgill.ca
Project title: Improving spatio-temporal image correlation spectroscopy (STICS) computational analysis platform
Spatio-temporal image correlation spectroscopy (STICS) is a fluorescence fluctuation analysis method to measure protein transport and interaction maps from analysis of input data sets from fluorescence microscopy image series of cells. STICS has been used to map dynamics involved in cell migration as well as signaling in cells. The current implementation of STICS is coded in MATLAB but does not have a graphical user interface and is not currently optimized for computational speed. The project would entail improving the STICS implementation via construction of a GUI based implementation, possible porting of the code to another more efficient language, parallelization of code or using GPU implementation to speed up the analysis of microscopy data sets.
Keywords: imaging, computer simulation, Image analysis and machine learning, cell signalling, cell migration
PI: Paul Wiseman
Department: Chemistry & Physics
email: paul.wiseman [at] mcgill.ca
Project title: Transcription factor networks in circulating monocytes during active viral lung infection
SARS-COV2 is a novel coronavirus which infects the airway and can lead to severe respiratory disfunction and disease called COVID-19. However, peripheral organ dysfunction has also been described. This is caused by a propagation of inflammation through the body by peripheral blood mononuclear cells (PBMCs). While the substantial effect this may have on patients is well known due to the COVID-19 pandemic, this is a feature common the viral infection in ICU patients. We propose to define the epigenetic character of PBMCs during the progression of viral disease to define the changes in genomic and transcriptional activation during disease. By studying the transcriptional network of the PBMCs we may gain insight into the signalling which the PBMCs are responding as well as the nature of their response. Through this, we may be able to develop better predictive models for ICU patients with viral pneumonia.
Keywords: genomics, machine learning
PI: Gregory Fonseca
Department: Medicine
email: gregory.fonseca [at] mcgill.ca
Project title: Proteomic analysis of GNAQ-mutant melanocytes and cancer-stem cells derived extracellular vesicles
Cutaneous melanoma and uveal melanoma are highly metastatic and deadly skin and eye cancer respectively. While both forms are developed from pigment-producing cells known as melanocytes, they are characterized by different driver mutations, and pathogenetic events remain unclear. Our first interest lies in the initiating GNAQ/11 mutations in uveal melanoma, as the mutations have been also found in over 80% of nevi cases which mostly remain benign. Epidemiological studies have implied an association of blue light exposure and the development of uveal melanoma. We would like to study proteomic alterations in GNAQ-mutant melanocytes following blue light exposure, which may help identify molecular signatures of tumorigenesis. Melanomas exhibit high tumor cell heterogeneity, which inspired our second interest in identifying subpopulations of tumor cells driving cancer metastasis. It has been proposed that subpopulations, termed cancer-stem cells, possessing both tumorigenic potential and stem-cell like features, promote drug-resistance and invasion to distant organ. Characterizing the proteomic profile of their secreted extracellular vesicles is vital to understand the crosstalk of cancer-stem cells with the microenvironment. In this project, we aim to identify key proteins involved in the tumorigenesis and metastasis of melanoma that may provide insights on the prevention and treatment of the diseases.
Keywords: extracellular vesicles, proteomics, melanoma, cancer stem cells, blue light, environmental exposure
PI: Julia Burnier
Department: Oncology and Pathology
email: Julia.burnier [at] mcgill.ca
Project title: Development of Agent-based simulation models for intervertebral disc biomaterials
lntervertebral disc (IVD) disorder is the leading cause of low back pain, which affects over 80% of the population worldwide. Direct medical costs per year were estimated as CAD $6-12 billion in Canada. Advanced IVD diseases often require surgical interventions such as spine fusion and total disc replacement. Clinical evidence has linked the current treatment options with post-surgical complications like altered spine mechanics and high long-term failure rates. High demand exists for new biomaterials and therapies to better instruct cellular activities and healing process. These conditions necessitate the proposed research on developing the computational platform to accelerate the invention and translation of novel bio-instructive biomaterials to improve the health and quality of life for millions of patients living with low back pain across the world.
Keywords: computer simulation, regenerative biomaterials
PI: Nicole Li-Jessen
Department: School of Communication Sciences and Disorders
email: nicole.li [at] mcgill.ca
Project title: Proteomic Characterization of Extracellular Vesicles (liquid biopsy) in a Rabbit Model of Melanoma
An estimated 85-95% of ocular melanomas arise in the uveal tract making uveal melanoma (UM) the prevalent intraocular malignancy in adults. Primary UM is often asymptomatic and easily treated, unfortunately 50% of patients form metastases plummeting survival rates to 15%. Given the statistics, there is an urgency to identify a plausible means by which these melanocytic tumors are disseminating. Recent studies have identified extracellular vesicles (EVs); biomolecule containing nanoparticles with the propensity to mediate metastasis through oncogenic reprogramming of target cells. EVs can be studied through a liquid biopsy - a minimally invasive approach to sample tumor material via the blood or other bio fluid. UM cell line derived EVs have been shown to induce oncogenic transformations in fibroblasts and resulted in tumor formation in a mouse model of metastasis. Further, our group has characterized the proteomic profile of EVs secreted from various UM cell lines and has begun analogous work with patient plasma derived EVs. The drawbacks in patient derived EVs is our inability to differentiate tumor originating EVs from those secreted for homeostatic reasons. Here we aim to simplify the issue through a rabbit model of UM, characterizing and separating EV proteomic profiles by species to identify potential biomarkers involved in metastatic UM.
Keywords: liquid biopsy, extracellular vesicles, cancer , proteomics, animal model
PI: Julia Burnier
Department: Oncology and Pathology
email: Julia.burnier [at] mcgill.ca
Project title: Decoding the role of calcium homeostasis in controlling the spiking of spinal inhibitory neurons
Injury-induced chronic neuropathic pain is debilitating and persists for years after tissue healing. Patients suffer from mechanical allodynia, a painful sensation in response to an innocuous stimulus. The allodynia is partially due to a loss of inhibition in the dorsal spinal cord. We demonstrated that a subset of inhibitory neurons, expressing the marker parvalbumin, are central to the development of allodynia. Indeed, their firing pattern changes after nerve injury to decrease their output. Single neuron sequencing data revealed the existence of several conductances in parvalbumin (PV) neurons. How these conductances shape the electrical properties of these neurons remains unclear.
Keywords: pain, Complex systems modeling, inhibition, firing pattern, ionic conductance, spinal cord, nerve injury, calcium homeostasis, dynamics data, machine classification
PI: Reza Sharif Naeini
Department: Physiology
email: reza.sharif [at] mcgill.ca
Project title: Modeling the role of Neuronal Endosomal Acid-Base Imbalances in synaptic plasticity and learning
For a long time, it was believed that the connections in the brain became fixed, and then simply faded. Research has shown that in fact the brain never stops changing through learning. Plasticity is the capacity of the brain to change with learning. Precise intracellular cargo trafficking is necessary for the development and plasticity of synapses within neuronal circuitry. Perturbations in this process can result in severe neurological deficits, such as the neurodevelopmental disorder Christianson syndrome (CS), autism and Alzheimer’s disease. Recently the SLC9A6 gene which encodes the electroneutral alkali cation (Na+ or K+)/proton (H+) exchanger isoform 6 (NHE6) found in recycling endosomes has been shown to play a critical role in this process. NHE6 tightly regulates graded acidification of compartments along the recycling and degradative pathways, which is critical for their biogenesis, as well as for cargo trafficking such as AMPA receptor delivery and cellular function. Interestingly, NHE6 has been shown to be down regulated in autism, CS and Alzheimer’s Disease resulting in problems in synaptic plasticity and neuronal circuits. We have developed an in house novel mouse model for loss of NHE6 which exhibits reduced learning and long term potentiation (LTP). Although we know which players are involved at the molecular level, from glutamate receptors to calcium binding proteins, we don't know how loss of NHE6 can modify the synapses susceptibility to LTP and the effect this will have on neuronal circuitry.
Keywords: computer simulation, neural circuits, memory, animal model, autism, neurodevelopmental disorders, electrophysiology
PI: Anne McKinney
Department: Pharmacology and Therapeutics
email: anne.mckinney [at] mcgill.ca
Project title: Gene set enrichment in amyotrophic lateral sclerosis
Amyotrophic lateral sclerosis (ALS) is a late-onset neurodegenerative disease with an incidence of 2-3 per 100,000 and an estimated lifetime risk of 1 in 472 and 1 in 350 in females and males, respectively. Initially, ALS symptoms consist of voluntary muscle weakness and atrophy leading to respiratory paralysis and death often within three-five years of diagnosis. Genetics is an important risk factor in ALS. In our recent study, we aggregated exome sequencing data from >5,000 ALS patients and >10,000 controls, to discover novel ALS risk factors. We showed a significant exome-wide enrichment of rare protein-truncating variants unique to ALS. This signal originated from constrained genes, which are a unique class of genes under strong purifying selection. Also, we identified a novel gene, DNAJC7, which is a constrained gene, mutated in patients (n=5,095) though completely absent in controls (n=28,910). Moreover, in a patient with a protein-truncating variant, p.Arg156Ter, we observed depleted DNAJC7 protein levels, further supporting the pathogenic role of DNAJC7 haploinsufficiency in ALS. Our approach of first interrogating a subset of genes based on their common biological function has led to a novel genetic discovery. We plan to leverage this same approach and extend it to multiple gene sets that may be enriched in neurodegeneration such as kinases, all heat-shock proteins, genes encoding RNA binding proteins, autophagy proteins, as well as other gene sets. If any gene passes the imposed statistical threshold, we will investigate whether patient cells are available in the Neuro labs or elsewhere, and we will work with our wet lab collaborators to perform the appropriate wet lab experiments based on the mechanism (gain or loss of function).
Keywords: statistical analyses, Big data analysis, Population genetics, Genomics, Gene discovery, bioinformatics, neuro
PI: Sali Farhan
Department: Neurosurgery and Neurology, and Human Genetics
email: sali.farhan [at] mcgill.ca
Project title: Large-scale multi-cohort prediction of psychopathology (ADHD, depression) using deep learning
Using deep learning methods, we seek to construct complex prediction models to identify patients at risk of depression/ADHD and determine which modifiable aspect of the environment should be targeted for maximizing improvement (personalized treatment). Our predictive models would consider both environmental and genetic risk/protective factors.
Keywords: deep learning, AI, risk prediction, depression, adhd, gxe
PI: Ashley Wazana
Department: Psychiatry
email: ashley.wazana [at] mcgill.ca
Population Health Research
Project title: >200k exomes to resolve the genomic overlap of social isolation with AD risk factors
Alzheimer’s disease and related dementias (ADRD) is a major public health burden – compounding over upcoming years due to longevity. Recently, epidemiological evidence hinted at the experience of social isolation in accelerating ADRD onset. In ~500,000 UK Biobank participants, we revisited traditional risk factors for developing ADRD in the context of loneliness and lacking social support. Across these measures of subjective and objective social deprivation, we have identified strong links between individuals’ social capital and various indicators of ADRD risk, which replicated across both population cohorts. The quality and quantity of daily social encounters may have deep connections with key aetiopathological factors, including stress management, cardiovascular effects on the heart-brain axis, sleep quality, substance use, and cognitive resource. Our population-scale findings call for urgent research efforts on the role of social lifestyle markers in ADRD neurodegeneration. As a next step, we are leveraging the open-access exomes of >200,000 UK Biobank participants to further identify genetic variation in the coding sequences of known ADRD genes as well as other psychiatric disease loci, and whether these genetic links can be replicated.
Keywords: None
PI: Danilo Bzdok
Department: Department of Biomedical Engineering
email: danilo.bzdok [at] mcgill.ca
Project title: Analyzing the field of Health Policy and Systems Research
The project aims to analyze the recent development of an important new field of health policy studies, Health Policy and Systems Research (HPSR). Starting around 2000, HPSR moved beyond the WHO, its institutional point of origin, to become an autonomous field with its own institutions and journals. Its goal was nothing less than to become a major influence on policy development in international agencies, national ministries of health, and local communities. Our goal in this project is to undertake a variety of computer mapping exercises in order to understand HPSR’s thematic and methodological development, as well as its relationship to the various other disciplines that cover some of the same ground. These exercises are of two types: scientometric analyses of the field’s literature and semantic analysis of more targeted corpuses of publications, such as the annual reports of the Alliance for Health Policy and Systems Research, the most important organization supporting HPSR research.
Keywords: None
PI:George Weisz
Department: Social Studies of Medicine
email: george.weisz [at] mcgill.ca
Project title: Prediction of chronic pain states in the aging population using genetic data
Chronic pain is a general public health problem affecting 20 % of the general population. The prevalence of chronic pain conditions reaches 50 % among older adults (>65 years), impacting on quality of life. Typically, chronic pain is defined as regular experience of pain at a given body site over a period of at least three months. SNP-chip heritability of chronic pain has been estimated at ~10%, and genetic correlations show a substantial overlap between chronic pain and psychiatric and immune-related phenotypes. Construction of polygenic risk scores (PRS) applied to chronic pain development would enable prediction of chronic pain states that could in turn be of clinical utility. We propose to draw on data from the Canadian Longitudinal Cohort on Aging as a primary study population (with other relevant study populations as target or replication populations). By considering pain phenotype data along with data on other conditions as trajectories over time, we seek statistical methods to construct PRS with higher levels of predictive power compared to PRS developed at a single timepoint.
Keywords: genomics, pain, Polygenic risk scores, GWAS, chronic disease, aging
PI: Audrey Grant
Department: Anesthesia
email: audrey.grant [at] mcgill.ca
