Biology subjects

Lau, Y.

Publications and source records attributed to Lau, Y..

4 recordsLinked to original sources

bioRxiv2023

Quantification of circulating tumor DNA using deep learning

Quantification of circulating tumor DNA (ctDNA) levels in blood enables non-invasive surveillance of cancer progression. Fragle is an ultra-fast deep learning-based method for ctDNA quantification directly from cell-free DNA fragment length profiles. We developed Fragle using low-pass whole genome sequence (lpWGS) data from multiple cancer types and healthy control cohorts, demonstrating high accuracy, and improved lower limit of detection in independent cohorts as compared to existing tumor-naive methods. Uniquely, Fragle is also compatible with targeted sequencing data, exhibiting high accuracy across both research and commercial targeted gene panels. We used this method to study longitudinal plasma samples from colorectal cancer patients, identifying strong concordance of ctDNA dynamics and treatment response. Furthermore, prediction of minimal residual disease in resected lung cancer patients demonstrated significant risk stratification beyond a tumor-naive gene panel. Overall, Fragle is a versatile, fast, and accurate method for ctDNA quantification with potential for broad clinical utility.

bioinformatics↗

bioRxiv2023

Drug repurposing for Leishmaniasis with Hyperbolic Graph Neural Networks

Leishmanisis, a neglected tropical disease caused by protozoan parasites of the genus Leishmania, affects millions of individuals living in poverty across the world and is second to malaria in parasitic causes of death. Although current drugs for treating leishmaniasis exist, these are either highly toxic, ineffective, or expensive. For this reason, there is an urgent need to identify affordable, safer, and more effective treatments. Drug repurposing is a promising method for identifying existing molecules with the potential to treat leishmaniasis. Here, we present a deep learning model for drug repurposing based on hyperbolic graph neural networks. We leverage experimentally validated protein-drug interactions and molecular descriptors across three different parasites to train and validate our model. The final network model shows significant gains over the best baseline model, with an 11.6% increase in precision of the top scoring 0.5% protein-drug pairs. Finally, our model identified two experimental drugs that could target three L. major proteins involved in drug resistance and cell cycle regulation, which play an essential role in ensuring the parasites survival inside the host. Author summaryLeishmanisis is a deadly and neglected tropical disease. Current treatments are ineffective, highly toxic, and unaffordable for the majority of affected individuals who live in extreme poverty. To accelerate the discovery and development of novel treatments against this disease, we develop a model that proposes existing drugs with the potential to treat leishmaniasis. Our methodology is based on hyperbolic graph neural networks, a class of deep learning models that can incorporate protein-drug interaction information from similar but well-studied parasites, in addition to utilizing chemical and molecular features. Empirically, this additional information leads to improved predictions of drug interactions with L. major proteins. Overall, the protein-drug pairs predicted by our model suggests two existing drugs that could target essential pathways for growth and survival of Leishmania parasites, which could be exploited either as a booster to increase the therapeutic effect of an existing anti-leishmaniasis drug, or as a novel chemotherapeutic treatment against the disease.

bioinformatics↗

bioRxiv2021

Heatrich-BS enables efficient CpG enrichment and highly scalable cell-free DNA methylation profiling

Genome wide analysis of cell-free DNA (cfDNA) methylation profile has been shown to be a promising approach for sensitive and specific multi-cancer detection. However, scaling these assays for clinical translation is impractical due to the high cost of whole genome bisulfite sequencing. We showed that the small fraction of GC-rich genome is highly enriched in CpG sites and disproportionately harbored the majority of cancer-specific methylation signature. Here, we report on the simple but effective Heat enrichment of CpG-rich regions for Bisulfite Sequencing (Heatrich-BS) platform that enables focused methylation profiling in these highly informative regions. Our novel method and bioinformatics algorithm enable high accuracy and sensitivity in tumor burden estimation and quantitative monitoring of colorectal patient response to treatment, at much reduced sequencing requirement. Heatrich-BS holds great potential for highly scalable screening and regular monitoring of cancer using liquid biopsy.

genomics↗

bioRxiv2020

A mechanism to prevent transformation of the Whi3 mnemon into a prion.

In response to deceptive courtship, budding yeast cells escape pheromone induced cell cycle arrest through coalescence of the G1/S inhibitor Whi3 into a dominant inactive super-assembly. Strikingly, Whi3 super-assemblies remain stable over many cell cycles in the mother cells and are not passed on to the daughter cells. Thereby, Whi3 coalescence encodes memory, conferring to it the property of a mnemon (Whi3mnem), a protein which conformational change maintain a trait that is permanent in the mother cell but is not inherited by daughter cells. Mnemons share structural features with prions, which are self-templating protein conformations that are inherited by daughter cells. Yet, how the maintenance and asymmetric inheritance of Whi3mnem are achieved is unknown. Here, we report that Whi3mnem is closely associated with endoplasmic reticulum (ER) membranes and retained in the mother cell by the presence of lateral membrane diffusion barriers at the bud neck. Strikingly, barrier defects made Whi3mnem propagate in a mitotically stable manner, like a prion. Alike Whi3mnem, transformation of Whi3 into a prion required its poly-glutamine prion-like domain. Thus, we propose that Whi3mnem is in a self-templating state, lending temporal stability to the memory that it encodes, while its anchorage into the compartmentalized membranes of the ER ensures its confinement in the mother cell and prevents its infectious propagation. These results suggest that confined self-templating super-assembly is a powerful mechanism for the long-term encoding of information.

cell biology↗