A5066417427- Genomics and Phylogenetic Studies
- vaccines and immunoinformatics approaches
- Genetic diversity and population structure
- SARS-CoV-2 and COVID-19 Research
- Gene expression and cancer classification
- Machine Learning in Bioinformatics
- COVID-19 epidemiological studies
- Electron and X-Ray Spectroscopy Techniques
- Visual perception and processing mechanisms
- Neural dynamics and brain function
- Machine Learning in Materials Science
- Neurobiology and Insect Physiology Research
- COVID-19 Clinical Research Studies
- Cellular Mechanics and Interactions
- Advanced Electron Microscopy Techniques and Applications
- CRISPR and Genetic Engineering
Massachusetts Institute of Technology
2023-2024
Bangladesh University of Engineering and Technology
2020-2024
Abstract The dynamic evolution of the severe acute respiratory syndrome coronavirus 2 virus is primarily driven by mutations in its genetic sequence, culminating emergence variants with increased capability to evade host immune responses. Accurate prediction such fundamental mitigating pandemic spread and developing effective control measures. This study introduces a robust interpretable deep-learning approach called PRIEST. innovative model leverages time-series viral sequences foresee...
Species tree estimation is frequently based on phylogenomic approaches that use multiple genes from throughout the genome. However, for a combination of reasons (ranging sampling biases to more biological causes, as in gene birth and loss), trees are often incomplete, meaning not all species interest have common set genes. Incomplete can potentially impact accuracy inference. We, first time, introduce problem imputing quartet distribution induced by incomplete trees, which involves adding...
Abstract Background Covid-19 pandemic, caused by the SARS-CoV-2 genome sequence of coronavirus, has affected millions people all over world and taken thousands lives. It is utmost importance that character this deadly virus be studied its nature analyzed. Methods We present here an analysis pipeline comprising a classification exercise to identify virulence sequences extraction important features from genetic material are used subsequently predict mutation at those interesting sites using...
Connectomics provides essential nanometer-resolution, synapse-level maps of neural circuits to understand brain activity and behavior. However, few researchers have access the high-throughput electron microscopes necessary generate enough data for whole circuit or reconstruction. To date, machine-learning methods been used after collection images by microscopy (EM) accelerate improve neuronal segmentation, synapse reconstruction other analysis. With computational improvements in processing...
We show how to improve the inference efficiency of an LLM by expanding it into a mixture sparse experts, where each expert is copy original weights, one-shot pruned for specific cluster input values. call this approach $\textit{Sparse Expansion}$. that, models such as Llama 2 70B, we increase number Sparse Expansion outperforms all other sparsification approaches same FLOP budget per token, and that gap grows sparsity increases, leading speedups. But why? To answer this, provide strong...
Abstract Leveraging retinotopic maps to parcellate the visual cortex into its respective sub-regions has long been a canonical approach characterizing functional organization of areas in mouse brain. However, with advent extensive connectomics datasets like MICrONS, we can now perform more granular analyses on biological neural networks, enabling us better characterize structural and profile cortex. In this work, propose statistical framework for analyzing MICrONS dataset, focusing our...
Abstract The dynamic evolution of the SARS-CoV-2 virus is largely driven by mutations in its genetic sequence, culminating emergence variants with increased capability to evade host immune responses. Accurate prediction such fundamental mitigating pandemic spread and developing effective control measures. In this study, we introduce a robust interpretable deep-learning approach called PRIEST. This innovative model leverages time-series viral sequences foresee potential mutations. Our...
This study leveraged the phylogenetic analysis of more than 10K strains novel coronavirus (SARS-CoV-2) from 67 countries. Due to requirement high-end computational power for analysis, we leverage a fast yet highly accurate alignment-free method develop tree out all coronavirus. K-Means clustering and PCA-based dimension reduction technique were used identify representative strain each location. The resulting was able highlight evolutionary relationships SARS-CoV-2 genome and, subsequently,...
Abstract Species tree estimation is frequently based on phylogenomic approaches that use multiple genes from throughout the genome. However, for a combination of reasons (ranging sampling biases to more biological causes, as in gene birth and loss), trees are often incomplete, meaning not all species interest have common set genes. Incomplete can potentially impact accuracy inference. We, first time, introduce problem imputing quartet distribution induced by incomplete trees, which involves...