ACMD Seminar: Applications in Machine Learning for Structure Elucidation

Jennifer Wei
Lead Software Engineer, OpenFold Project

Tuesday, December 12, 2023, 3:00-4:00 PM ET (1:00-2:00 PM MT)

In person at: Gaithersburg Bldg. 101 LR D with VTC to Boulder 1-4072
Online at: ZoomGov

Add this talk to your calendar: https://inet.nist.gov/calendar/ics/2272701

Abstract: Mass spectrometry is an important analytical tool used to characterize unknown samples. Several different algorithms have been developed to identify compounds given an electron ionization mass spectra, either by directly predicting the molecule given the spectra itself, or by augmenting mass spectral libraries with predicted spectra. In the first part of my talk, I will discuss some of these algorithms, and ways in which machine learning has been applied to augment these tasks.

In the second part, I will discuss the OpenFold Project, whose mission is to build open source software for biology and drug discovery. Last year, our team released an open source, fully trainable reproduction of the AlphaFold model. This OpenFold model fully replicated the accuracy of the original model, and was optimized for computational efficiency on GPUs. I will also share some of our current work in new model features, and next steps for the OpenFold project.

Bio: Jennifer Wei is a specialist in the applications of machine learning to the chemical sciences. In her PhD studies in Chemical Physics at Harvard University, she worked on machine learning algorithms for organic reaction prediction, generative models for molecules, and mass spectrometry prediction. She later joined Google Brain Research and worked on mapping olfactory properties of small molecules, discovering new mosquito repellents, and functional prediction of nanobody libraries produced from alpacas. Jennifer is now lead software engineer on the OpenFold Project, building open source machine learning models for protein structure prediction.

Host: Tony Kearsley

Note: This talk will be recorded to provide access to NIST staff and associates who could not be present to the time of the seminar. The recording will be made available in the Math channel on NISTube, which is accessible only on the NIST internal network. This recording could be released to the public through a Freedom of Information Act (FOIA) request. Do not discuss or visually present any sensitive (CUI/PII/BII) material. Ensure that no inappropriate material or any minors are contained within the background of any recording. (To facilitate this, we request that cameras of attendees are muted except when asking questions.)

Note: Visitors from outside NIST must contact Meliza Lane at least 24 hours in advance.