University of Tennessee, Knoxville TRACE: Tennessee Research and Creative Exchange Doctoral Dissertations Graduate School 8-2015 QUANTITATIVE AND FUNCTIONAL ANALYSIS PIPELINE FOR LABEL-FREE METAPROTEOMICS DATA AND ITS APPLICATIONS Lang Ho Lee University of Tennessee - Knoxville, [email protected] Follow this and additional works at: https://trace.tennessee.edu/utk_graddiss Part of the Biochemistry Commons, Bioinformatics Commons, and the Integrative Biology Commons Recommended Citation Lee, Lang Ho, "QUANTITATIVE AND FUNCTIONAL ANALYSIS PIPELINE FOR LABEL-FREE METAPROTEOMICS DATA AND ITS APPLICATIONS. " PhD diss., University of Tennessee, 2015. https://trace.tennessee.edu/utk_graddiss/3435 This Dissertation is brought to you for free and open access by the Graduate School at TRACE: Tennessee Research and Creative Exchange. It has been accepted for inclusion in Doctoral Dissertations by an authorized administrator of TRACE: Tennessee Research and Creative Exchange. For more information, please contact [email protected]. To the Graduate Council: I am submitting herewith a dissertation written by Lang Ho Lee entitled "QUANTITATIVE AND FUNCTIONAL ANALYSIS PIPELINE FOR LABEL-FREE METAPROTEOMICS DATA AND ITS APPLICATIONS." I have examined the final electronic copy of this dissertation for form and content and recommend that it be accepted in partial fulfillment of the equirr ements for the degree of Doctor of Philosophy, with a major in Life Sciences. Nathan C. VerBerkmoes, Tim E. Sparer, Major Professor We have read this dissertation and recommend its acceptance: Tamah Fridman, Arnold M. Saxton, Chongle Pan Accepted for the Council: Carolyn R. Hodges Vice Provost and Dean of the Graduate School (Original signatures are on file with official studentecor r ds.) QUANTITATIVE AND FUNCTIONAL ANALYSIS PIPELINE FOR LABEL-FREE METAPROTEOMICS DATA AND ITS APPLICATIONS A Dissertation Presented for the Doctor of Philosophy Degree The University of Tennessee, Knoxville Lang Ho Lee August 2015 Copyright © 2015 by Lang Ho Lee All rights reserved. ii ACKNOWLEDGEMENTS First, I’d like to gratefully and sincerely thank Dr. Nathan VerBerkmoes for guiding me through the completion of my dissertation and subsequent Ph.D. He gave me a chance to work at Oak Ridge National Laboratory as well as Berg Phama Inc. as a research assistant. Also, he never stopped supporting me while he had no easy matters. I believe that he will successfully continue his researches at his new position. Secondly, I’m very grateful to the remaining members of my dissertation committee, Dr. Arnold Saxton, Dr. Tim Sparer, Dr. Chongle Pan and Dr. Tamah Fridman. Their academic support and input are greatly appreciated. Especially, I appreciate to Dr. Arnold Saxton who has provided me another home as well as academic advises. I’d also like to express my gratitude to Dr. Steven Wilhelm and Dr. Bhanu Rekepalli for their support and help. Thirdly, I’d like to thank my parents Jae Duk Lee and Ok Hee Kim and parents in law Young Gak Shin and Jung Bae Kwak as well as my relatives. Their prayers and supports have helped me a lot. I couldn’t finish Ph.D. without their support and encouragement. Finally, and most importantly, my upmost gratitude goes to my family, Bo Hee and Yoel. They followed after me and moved to where they never heard before. Their encouragement, support and patience were undeniably precious source of power to keep on studying. Obviously, without these two, it would not iii have been possible to complete my study and to build up all things together. Above all, I am greatly thankful to God who guided, inspired and encouraged me to go this way. iv ABSTRACT Since the large-scale metaproteome was first reported in 2005, metaproteomics has advanced at a tremendous rate both in its quantitative and qualitative metrics. Furthermore metaproteomics is now being applied as a general tool in microbial ecology in a large variety of environmental studies. Though metaproteomics is becoming a useful and even a standard tool for the microbial ecologist, standardized bioinformatics pipelines are not readily available. Therefore, we developed quantitative and functional analysis pipeline for metaproteomics (QFAM) to help analyze large and complicated metaproteomics data in a robust and timely fashion with outputs designed to be simple and clearly understood by the microbial ecologist. QFAM starts by running peptide-spectrum searches against resultant MS/MS datasets with mixed metagenome/appropriate protein FASTA database. Its primary search algorithm is MyriMatch/IDPicker. MyriMatch/IDPicker uses multi-CPUs effectively, has an accurate scoring-system, correctly use the high MS accuracy data, and finally has a robust method for protein determination. These are required features for metaproteomics requiring large protein database and complicated peptide-structure. QFAM has quantitative (QAM) and functional (FAM) analysis to provide dependable protein signatures and confident information for understanding the characteristics of the metaproteome. QAM employs a ’selfea’ R package, which v provides probability models as well as Cohen’s effect sizes. Our benchmark data test and Monte Carlo simulation results show that selfea can reduce false positives efficiently while losing few true positives; one of the key goals of proteomics and/or metaproteomics experiments. FAM has two modules: BioSystems and COG analysis. The BioSystems module is most appropriate for well-annotated model organisms, such as humans, whereas the COG module is useful for less-annotated microorganisms and metagenome sequences. Both modules provide an enrichment test using Fisher’s exact-test and a significance test using selfea. With two statistics, FAM generates differentially enriched functional terms that are insightful for discerning biological information held behind the metaproteome data. Two application studies in chapter 4 and 5 show how QFAM can be employed for metaproteomics data analysis. QFAM is distinguished from other proteomics pipelines by multiprocessing as well as quantitative and functional analysis. vi TABLE OF CONTENTS CHAPTER I INTRODUCTION .............................................................................. 1 Metaproteomics ................................................................................................. 1 Metaproteomics in proteomics ....................................................................... 1 Shotgun metaproteomics ............................................................................... 4 LC-MS/MS analysis ..................................................................................... 15 Bioinformatics for metaproteomics .................................................................. 28 Peptide-spectrum search and identification ................................................. 28 Post-process of PSM search results ............................................................ 41 Quantitative and functional analysis pipeline for metaproteomics ............... 51 CHAPTER II QUANTITATIVE AND FUNCTIONAL ANALYSIS PIPELINE FOR METAPROTEOMICS DATA ............................................................................... 56 Quantitative and functional analysis for metaproteomics (QFAM) .................. 56 Objectives of QFAM ..................................................................................... 56 General information of QFAM ...................................................................... 56 Module for LC-MS/MS data analysis ............................................................... 59 Automated module for DB search and protein assembly ............................. 59 Making contrast file, comparing MS runs, and normalization ...................... 62 Module for quantitative and functional analysis ............................................... 65 Quantitative analysis module (QAM) ........................................................... 65 Functional analysis module (FAM) ............................................................... 69 BioSystems analysis of FAM ....................................................................... 71 COG analysis of FAM .................................................................................. 76 Other modules of QFAM .............................................................................. 83 Conclusion ....................................................................................................... 89 CHAPTER III Selfea: R package for reliable feature selection ........................... 93 Introduction ...................................................................................................... 93 Methods ........................................................................................................... 96 vii Label free LC-MS/MS Proteomics Data ....................................................... 96 Calculation of Cohen’s effect sizes and statistics ........................................ 96 Benchmark data test and fecal proteome data test ..................................... 98 Monte Carlo simulation ................................................................................ 99 User guide for selfea package ................................................................... 100 Algorithm ....................................................................................................... 100 Benchmark test .............................................................................................. 104 Tests using Gregori