Ping Yang

219 Goessmann Laboratory

686 N Pleasant St

Amherst, MA 01003

I am a third year PhD student in Chemical Engineering, University of Massachusetts Amherst. I am working with Dr. Peng Bai as a graduate research assistant. My major research lies in leveraging machine learning to guide property prediction of materials, especially Graph Neural Networks as molecules inherent have graph structures.

Prior to coming to UMass, I obtained my master degree from Oregon State University, where I work with Prof. Simon in The Simon Research Ensemble. We developed a machine learning model to predict the toxicity of pesticides to honey bees via Support Vector Machines with Random Walk graph kernels. I hold bachelor degree in Chemical Engineering from East China University of Science and Technology.

I was born and raised in Wuhan, China, a famous city now as the well-known reason. In my free time, I enjoying playing video games and spending time with friends on outdoor activities or board games.

news

Dec 22, 2022	Officially joined Bai Lab.
Jul 05, 2022	Bee toxicity paper on the cover of JCP and in Scilight
Apr 15, 2022	Ping Yang is officially headed to UMass Amherst in the fall to pursue his PhD in Chemical Engineering!

latest posts

Sep 06, 2023	Markdown Grammar
May 22, 2021	a distill-style blog post

selected publications

Classifying the Toxicity of Pesticides to Honey Bees via Support Vector Machines with Random Walk Graph Kernels

Ping Yang, E. Adrian Henle, Xiaoli Z. Fern, and Cory M. Simon

J. Chem. Phys., 2022

Abs HTML

Pesticides benefit agriculture by increasing crop yield, quality, and security. However, pesticides may inadvertently harm bees, which are valuable as pollinators. Thus, candidate pesticides in development pipelines must be assessed for toxicity to bees. Leveraging a dataset of 382 molecules with toxicity labels from honey bee exposure experiments, we train a support vector machine (SVM) to predict the toxicity of pesticides to honey bees. We compare two representations of the pesticide molecules: (i) a random walk feature vector listing counts of lengthL walks on the molecular graph with each vertex- and edge-label sequence and (ii) the Molecular ACCess System (MACCS) structural key fingerprint (FP), a bit vector indicating the presence/absence of a list of pre-defined subgraph patterns in the molecular graph. We explicitly construct the MACCS FPs but rely on the fixed-length-L random walk graph kernel (RWGK) in place of the dot product for the random walk representation. The L-RWGK-SVM achieves an accuracy, precision, recall, and F1 score (mean over 2000 runs) of 0.81, 0.68, 0.71, and 0.69, respectively, on the test data set—with L = 4 being the mode optimal walk length. The MACCS-FP-SVM performs on par/marginally better than the L-RWGK-SVM, lends more interpretability, but varies more in performance. We interpret the MACCS-FP-SVM by illuminating which subgraph patterns in the molecules tend to strongly push them toward the toxic/non-toxic side of the separating hyperplane.