Work

Differential abundance analyses of human microbiota in Parkinson’s disease

Year: 2019
Type: dissertation
Study design: Observational Study
Abstract: Parkinson's disease is the second most common neurodegenerative disease in the world, and in spite of decades of research, the cause of the non-familial form of the disease is not known. There are currently no medications to slow down the progression of the disease nor good biomarkers for early diagnosis, even though the earliest non-motor symptoms can appear years or even decades before the onset of motor symptoms. The microbial inhabitants of the human body have recently been implicated in various medical conditions, including neurodegenerative disorders. They could offer new insight into the pathogenesis of Parkinson's disease, particularly since a microbial agent has long been suspected to play a part in the process. A typical question in studies surveying human microbiota is which specific microbial taxa differ between groups of interest, such as patients with a disease and control subjects. There are many statistical tools for performing these analyses, also known as differential abundance comparisons. The aim of my doctoral thesis was to explore the potential associations of Parkinson's disease and human microbiota, particularly of the mouth, the nose and the gut, with an additional focus on the statistical tools used for comparing differentially abundant bacterial taxa. All four publications included in my thesis were based on samples from the same subjects: 76 patients with Parkinson's disease and 76 control subjects with no signs of parkinsonism. The studies also used the same methodology, 16S rRNA gene amplicon sequencing, to compare bacteria from oral and nasal swab samples and fecal samples of these subjects. Over the course of the four publications and in a previously unpublished analysis, I compared differentially abundant taxa with six tools: Metastats, LEfSe, metagenomeSeq, DESeq2, ANCOM, and an approach based on random forests. Our results suggested that the bacterial communities of the gut and the mouth differ between Parkinson's patients and control subjects, with statistically significant differences in beta diversity and in the abundances of several bacterial taxa. Differences in gut microbiota could also be detected at a follow-up time point with samples collected two years after the initial sampling. Additionally, there were differences between the gut bacteria of Parkinson's patients with and without irritable bowel syndrome -like symptoms. For nasal bacteria, there were no differences between the patient and control groups in diversity nor the amounts of specific bacteria. Regarding the differential abundance analyses, comparing gut bacteria of patients and controls from the same samples with six different tools highlighted the wide variation in the lists of significant results, which often did not overlap except for a handful of taxa. While a few benchmarking studies have previously contrasted some of the tools, there is a definite need for further standardized testing to guide researchers in choosing between them. Despite these discrepancies, all tools tested in this thesis supported Parkinson's patients having a decreased abundance of the family Prevotellaceae in their gut. This difference in abundance could also be detected at the follow-up time point. As several other research groups have reported seeing a decrease in Prevotellaceae after our pilot publication, it is emerging as one of the key changes in microbiota associated with Parkinson's disease.
Language: English
Open Access status: green
OpenAlex ID: W2940826496
Cites: 341
Cited by:
Related to: 9