PERFect: PERmutation Filtering test for microbiome data

Abstract

The human microbiota composition is associated with a number of diseases including obesity, inflammatory bowel disease, and bacterial vaginosis. Thus, microbiome research has the potential to reshape clinical and therapeutic approaches. However, raw microbiome count data require careful pre-processing steps that take into account both the sparsity of counts and the large number of taxa that are being measured. Filtering is defined as removing taxa that are present in a small number of samples and have small counts in the samples where they are observed. Despite progress in the number and quality of filtering approaches, there is no consensus on filtering standards and quality assessment. This can adversely affect downstream analyses and reproducibility of results across platforms and software. We introduce PERFect, a novel permutation filtering approach designed to address two unsolved problems in microbiome data processing: (i) define and quantify loss due to filtering by implementing thresholds and (ii) introduce and evaluate a permutation test for filtering loss to provide a measure of excessive filtering. Methods are assessed on three "mock experiment" data sets, where the true taxa compositions are known, and are applied to two publicly available real microbiome data sets. The method correctly removes contaminant taxa in "mock" data sets, quantifies and visualizes the corresponding filtering loss, providing a uniform data-driven filtering criteria for real microbiome data sets. In real data analyses PERFect tends to remove more taxa than existing approaches; this likely happens because the method is based on an explicit loss function, uses statistically principled testing, and takes into account correlation between taxa. The PERFect software is freely available at https://github.com/katiasmirn/PERFect.

Response of the human gut and saliva microbiome to urbanization in Cameroon

“…Filtering for beta diversity analysis and community type determination. In order to get rid of the noise and to reduce the complexity of the dataset, we performed a two-step unsupervised filtering implemented in the PERFect package 73 . Briefly, the method compares the total covariance of the community data ("OTU table") before and after removal of a taxon.…”

Effects of microbiome rare taxa filtering on statistical analysis

“…It extends traditional rule of thumb filtering approaches to find the best subset of retained taxa for further analysis by implementing statistical data-driven significance cut-off thresholds. The current method for such approach is PERFect, a principled filtering test that removes taxa with insignificant contribution to the total covariance [14]. Specifically, this method ranks taxa importance, measures their contribution to the total covariance, and quantifies the chance that the loss increases for a set of filtered taxa is due to randomness using permutation tests.…”

“…Most filtering approaches are based on the rules of thumb, which vary from lab-to-lab. Recently, a filtering loss measure and a principled filtering test, namely PERFect [14], is introduced for deciding which taxa to remove. These methods are implemented in Bioconductor package PERFect [15], which includes a novel fast implementation of the permutation PERFect method.…”

“…Recently, a filtering loss measure and a principled filtering test, namely PERFect [14], is introduced for deciding which taxa to remove. These methods are implemented in Bioconductor package PERFect [15], which includes a novel fast implementation of the permutation PERFect method. The implemented approach successfully reduces the original algorithm running time by almost four times.…”

“…Traditional filtering methods were previously compared to the PERFect approach proposed by Smirnova et al [14] and tested on two datasets acquired from mock community experiments carried out at Virginia Commonwealth University (VCU) [12,18] and a reagent and laboratory contamination dataset [17]. The authors used the number of contaminant taxa removed from the mock datasets as the method evaluation criteria.…”