Joseph Szymborski

Joseph Szymborski

Auto-generated profile, based on publicly available information

Joseph Szymborski

Department of Electrical and Computer Engineering, McGill University, Montreal, Canada

Research Focus & Metrics

Academic Career Length

11Years (2016-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
23 May 2018 in Oncogene
Verified
Journal Article
Open access
PDF file
SHSara Al Habyan
Sara Al Habyan
CKChristina Kalos
Christina Kalos
JSJoseph Szymborski
Joseph Szymborski
LMLuke McCaffrey
Luke McCaffrey
Ovarian cancer is the most lethal gynecological cancer, where survival rates have had modest improvement over the last 30 years. Metastasis of cancer cells is a major clinical problem, and patient mortality occurs when ovarian cancer cells spread beyond the confinement of ovaries. Disseminated ovarian cancer cells typically spread within the abdomen, where ascites accumulation aids in their transit. Metastatic ascites contain multicellular spheroids, which promote chemo-resistance and recurrence. However, little is known about the origin and mechanisms through which spheroids arise. Using live-imaging of 3D culture models and animal models, we report that epithelial ovarian cancer (EOC) cells, the most common type of ovarian cancer, can spontaneously detach as either single cells or clusters. We report that clusters are more resistant to anoikis and have a potent survival advantage over single cells. Using in vivo lineage tracing, we found that multicellular spheroids arise preferentially from collective detachment, rather than aggregation in the abdomen. Finally, we report that multicellular spheroids from collective detachment are capable of seeding intra-abdominal metastases that retain intra-tumoral heterogeneity from the primary tumor.

1 August 2017 in Genes & Development
Verified
Journal Article
Open access
PDF file
RHRuba Halaoui
Ruba Halaoui
CRCarlis Rejon
Carlis Rejon
SCSudipa June Chatterjee
Sudipa June Chatterjee
JSJoseph Szymborski
Joseph Szymborski
SMSarkis Meterissian
Sarkis Meterissian
LMLuke McCaffrey
Luke McCaffrey
Epithelial cancers (carcinoma) account for 80%–90% of all cancers. The development of carcinoma is associated with disrupted epithelial organization and solid ductal structures. The mechanisms underlying the morphological development of carcinoma are poorly understood, but it is thought that loss of cell polarity is an early event. Here we report the characterization of the development of human breast lesions leading to carcinoma. We identified a unique mechanism that generates solid ducts in carcinoma through progressive loss of polarity and collapse of the luminal architecture. This program initiates with asymmetric divisions of polarized cells that generate a stratified epithelium containing both polarized and depolarized cells. Stratified regions form cords that penetrate into the lumen, subdividing it into polarized secondary lumina. The secondary lumina then collapse with a concomitant decrease in RhoA and myosin II activity at the apical membrane and ultimately lose apical–basal polarity. By restoring RhoA activity in mice, ducts maintained lumen and cell polarity. Notably, disrupted tissue architecture through luminal collapse was reversible, and ducts with a lumen were re-established after oncogene suppression in vivo. This reveals a novel and common mechanism that contributes to carcinoma development by progressively disrupting cell and tissue organization.

30 June 2022 in Bioinformatics
Verified
Journal Article
Closed access
PDF file
JSJoseph Szymborski
Joseph Szymborski
AEAmin Emad
Amin Emad
Motivation: Computational methods for the prediction of protein–protein interactions (PPIs), while important tools for researchers, are plagued by challenges in generalizing to unseen proteins. Datasets used for modelling protein–protein predictions are particularly predisposed to information leakage and sampling biases.Results: In this study, we introduce RAPPPID, a method for the Regularized Automatic Prediction of Protein–Protein Interactions using Deep Learning. RAPPPID is a twin Averaged Weight-Dropped Long Short-Term memory network which employs multiple regularization methods during training time to learn generalized weights. Testing on stringent interaction datasets composed of proteins not seen during training, RAPPPID outperforms state-of-the-art methods. Further experiments show that RAPPPID’s performance holds regardless of the particular proteins in the testing set and its performance is higher for experimentally supported edges. This study serves to demonstrate that appropriate regularization is an important component of overcoming the challenges of creating models for PPI prediction that generalize to unseen proteins. Additionally, as part of this study, we provide datasets corresponding to several data splits of various strictness, in order to facilitate assessment of PPI reconstruction methods by others in the future.Availability and implementation: Code and datasets are freely available at https://github.com/jszym/rapppid and Zenodo.org.Supplementary information: Supplementary data are available at Bioinformatics online.

25 July 2024 in Briefings in Bioinformatics
Verified
Journal Article
Open access
PDF file
JSJoseph Szymborski
Joseph Szymborski
AEAmin Emad
Amin Emad
An overwhelming majority of protein–protein interaction (PPI) studies are conducted in a select few model organisms largely due to constraints in time and cost of the associated ‘wet lab’ experiments. In silico PPI inference methods are ideal tools to overcome these limitations, but often struggle with cross-species predictions. We present INTREPPPID, a method that incorporates orthology data using a new ‘quintuplet’ neural network, which is constructed with five parallel encoders with shared parameters. INTREPPPID incorporates both a PPI classification task and an orthologous locality task. The latter learns embeddings of orthologues that have small Euclidean distances between them and large distances between embeddings of all other proteins. INTREPPPID outperforms all other leading PPI inference methods tested on both the intraspecies and cross-species tasks using strict evaluation datasets. We show that INTREPPPID’s orthologous locality loss increases performance because of the biological relevance of the orthologue data and not due to some other specious aspect of the architecture. Finally, we introduce PPI.bio and PPI Origami, a web server interface for INTREPPPID and a software tool for creating strict evaluation datasets, respectively. Together, these two initiatives aim to make both the use and development of PPI inference tools more accessible to the community.

13 February 2026 in Nature Machine Intelligence
Verified
Journal Article
Open access
PDF file
JSJoseph Szymborski
Joseph Szymborski
AEAmin Emad
Amin Emad
With the growing pervasiveness of pretrained protein language models (pLMs), pLM-based methods are increasingly being put forward for the protein–protein interaction (PPI) inference task. Here we identify and confirm that existing pretrained pLMs are a source of data leakage for the downstream PPI task. We characterize the extent of the data leakage problem by training and comparing small and efficient pLMs on a dataset that controls for data leakage (strict) with one that does not (non-strict). Although data leakage from pretrained pLMs cause a measurable inflation of testing scores, we find that this does not necessarily extend to other, non-paired biological tasks such as protein keyword annotation. Further, we find no connection between the context lengths of pLMs and the performance of pLM-based PPI inference methods on proteins with sequence lengths that surpass it. Furthermore, we show that pLM-based and non-pLM-based models fail to generalize in tasks such as prediction of the human-SARS-CoV-2 PPIs or the effect of point mutations on binding affinities. This study demonstrates the importance of extending existing protocols for the evaluation of pLM-based models applied to paired biological datasets and identifies areas of weakness of current pLM models.

12 August 2025 in bioRxiv
Verified
Preprint
Open access
NBNicolas Bolouri
Nicolas Bolouri
JSJoseph Szymborski
Joseph Szymborski
AEAmin Emad
Amin Emad
Effectively integrating data modalities pertaining to proteins’ amino acid sequences, three-dimensional structures, and curated text-based descriptions of their biochemical and functional properties can lead to informative representations capturing different views of proteins. Here, we introduce CLASP, a unified tri-modal framework that combines the strengths of geometric deep learning, natural large language models (LLMs), protein language models (pLMs), and contrastive learning to learn informative protein representations based on their structure, amino acid sequence, and text-based biochemical and functional descriptions. We show that CLASP enables accurate zero-shot classification and retrieval tasks, such as matching a protein structure to its sequence or description, outperforming state-of-the-art baselines. CLASP embeddings also exhibit superior clustering by protein family, and ablation studies confirm that all three modalities contribute synergistically to performance. Our results highlight the power of integrating structural, sequential, and textual signals in a single model, establishing CLASP as a general-purpose embedding framework for protein understanding.