Using Machine Learning Algorithms to Predict Conditions for Protein Crystallization Öffentlichkeit

Abstract

Radical S-adenosyl-methionine (rSAM) enzymes comprise a large, primarily uncharacterized metalloenzyme superfamily, important for the biosynthesis of a wide range of natural products across many living organisms. All constituent members rely on the reductive cleavage of SAM bound to a [4Fe- 4S] cluster to generate a highly reactive 5'-deoxyadenosyl radical. This radical mechanism facilitates challenging chemistries and enables a diverse array of reactivities including methylation and isomerization of substrates. However, little is known about the structural basis for this impressive breadth of reaction outcomes. Protein crystallography is a powerful method for determining the 3-dimensional structure of proteins, but the logistical complications of this method, coupled with rSAM enzyme sensitivity to molecular oxygen complicate crystallization of enzymes in this superfamily. Correspondingly, structures are rare.

Using machine learning methods to predict the conditions in which a crystal will grow for a given protein would greatly increase the efficiency of protein crystallography and aid in developing a deeper understanding of the rSAM family. This goal was partitioned into two aims: (1) Developing a clear understanding of current protein crystallography methodology by crystallizing the rSAM enzyme YydG. (2) Developing machine learning algorithms to predict crystallization conditions for the rSAM enzyme SuiB. After aim one was met and sufficient expertise was gained with protein crystallography through working with YydG, three machine learning models were applied to crystallization data for SuiB. The models all performed higher than 50% accuracy, indicating that a computational approach to predicting crystallization conditions can improve the process of developing protein crystals. 

Table of Contents

Introduction...........................................................................................pg. 1 

Aim 1 Background.............................................................................. ....pg. 6 

Aim 1 Methods.................................................................................. ....pg. 7 

Aim 1 Results.................................................................................... ....pg. 9 

Aim 2 Background.............................................................................. ....pg. 12 

Aim 2 Methods.................................................................................. ....pg. 15 

Aim 2 Results.................................................................................... ....pg. 19 

Future Directions.....................................................................................pg. 25 

References.............................................................................................pg. 29 

About this Honors Thesis

Rights statement

• Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.

School	Emory College
Department	Chemistry
Degree	B.S.
Submission	Honors Thesis
Language	English
Research Field	Computer Science Chemistry, Biochemistry
Stichwort	Protein Crystallography Machine Learning rSAM Enzymes
Committee Chair / Thesis Advisor	Dr. Katherine Davis, Emory University
Committee Members	Dr. Joel Bowman, Emory University Dr. James Kindt, Emory University

Zuletzt geändert

Primary PDF

Thumbnail	Title	Date Uploaded	Actions
	Using Machine Learning Algorithms to Predict Conditions for Protein Crystallization ()	2022-04-29 12:09:10 -0400	Press to Select an action Download

Supplemental Files

Thumbnail	Title	Date Uploaded	Actions