Contact UsComputer Science Department
500 College Avenue
Swarthmore, PA 19081
Email: info at cs.swarthmore.edu
Copyright 2009 Swarthmore College. All rights reserved.
Ameet Soni: Probabilistic Methods for Biomedical Applications
My research falls into two general categories:
- Probabilistic Methods in machine learning, specifically learning and inference in complex graphical models such as in statistical relational learning
- Biomedical Applications including clinical diagnosis, protein-structure prediction, gene modeling, biomedical-image analysis, and information extraction from biomedical texts
Recently, I worked on ACMI (Automated Crystallographic Map Interpretation). The task of determining protein structures has been a central one to the biological community for several decades. The structure allows biologists to extract information about the underlying biology of a protein, and has implications for various applications such as disease treatment, drug design, and protein design. The main objective of the project was:
Given the electron-density map (3D image) and a primary sequence of a target protein, produce a three- dimensional, physically-feasible, all-atom model of the target protein's structure.
This is a difficult task with low-resolution images as they number of variables is computationally difficult to handle, coupled with poor noisy images. Our group developed ACMI, a probalistic technique for determining protein structures. To reason over this difficult search space, ACMI utilizes Markov Random Fields - a type of probabilistic model - in combination with sophisticated inference techniques to produce complete and accurate structures for previously unsolvable protein structures.
More recently, I have begun researching several different computational problems in the area of MRI braining imaging. Specifically, I seek to apply probabilistic methods to improve the ability to diagnose onset of Alzheimer's disease and other cognitive impairments. Currently, my group is pursuing methods for segmenting anatomical regions of the brain using graph-based methods as well intuit anatomical relationships between patients in the larger ADNI study.
Ameet Soni's homepage: https://www.cs.swarthmore.edu/~soni