Integrating phylogenetic and functional data in microbiome studies

PhyloFunc: phylogeny-informed functional distance as a new ecological metric for metaproteomic data analysis

BACKGROUND

Beta-diversity is a fundamental ecological metric for exploring dissimilarities between microbial communities. On the functional dimension, metaproteomics data can be used to quantify beta-diversity to understand how microbial community functional profiles vary under different environmental conditions. Conventional approaches to metaproteomic functional beta-diversity often treat protein functions as independent features, ignoring the evolutionary relationships among microbial taxa from which different proteins originate. A more informative functional distance metric that incorporates evolutionary relatedness is needed to better understand microbiome functional dissimilarities.

RESULTS

Here, we introduce PhyloFunc, a novel functional beta-diversity metric that incorporates microbiome phylogeny to inform on metaproteomic functional distance. Leveraging the phylogenetic framework of weighted UniFrac distance, PhyloFunc innovatively utilizes branch lengths to weigh between-sample functional distances for each taxon, rather than differences in taxonomic abundance as in weighted UniFrac. Proof of concept using a simulated toy dataset and a real dataset from mouse inoculated with a synthetic gut microbiome and fed different diets show that PhyloFunc successfully captured functional compensatory effects between phylogenetically related taxa. We further tested a third dataset of complex human gut microbiomes treated with five different drugs to compare PhyloFunc's performance with other traditional distance methods. PCoA and machine learning-based classification algorithms revealed higher sensitivity of PhyloFunc in microbiome responses to paracetamol. We provide PhyloFunc as an open-source Python package (available at https://pypi.org/project/phylofunc/ ), enabling efficient calculation of functional beta-diversity distances between a pair of samples or the generation of a distance matrix for all samples within a dataset.

CONCLUSIONS

Unlike traditional approaches that consider metaproteomics features as independent and unrelated, PhyloFunc acknowledges the role of phylogenetic context in shaping the functional landscape in metaproteomes. In particular, we report that PhyloFunc accounts for the functional compensatory effect of taxonomically related species. Its effectiveness, ecological relevance, and enhanced sensitivity in distinguishing group variations are demonstrated through the specific applications presented in this study. Video Abstract.

Influence of Saccharomyces cerevisiae CNCM I-1077 on the fecal pH, markers of gut permeability, fecal microbiota, and markers of systemic inflammation in sedentary horses fed a high-starch diet

Thirty mature Quarter Horse geldings were used in a completely randomized 32-d study to test the hypotheses that supplemental live Saccharomyces cerevisiaeCNCM; I-1077 improves apparent digestion, stabilizes the fecal pH, reduces gut permeability, maintains microbial communities, and decreases inflammation in horses fed a high-starch diet. Horses were stratified by body weight (BW), age, and body condition score (BCS) to one of two treatments: concentrate formulated with 2 g starch · kg BW-1 · meal-1 (control (CON); n = 15) or the same concentrate top-dressed with 25 g/d S. cerevisiae CNCM I-1077 (treatment (SC); n = 15; 8 × 108CFU). Horses were fed individually in stalls every 12 h. Between meals, horses were housed in dry lots with ad libitum access to water and Coastal bermudagrass hay. On days 0 and 32, BW and BCS were recorded, and blood was collected before feeding and 2, 8, 16, and 24 h postmeal on day 32 to analyze serum D-lactate. Fecal samples were collected on days 0, 16, and 32 at 8, 16, and 24 h postmeal for fecal pH and starch content. Intake and fecal production were recorded over 4 d to measure digestibility on days 28-31. Whole blood total bacterial counts and 16S fecal microbiota rRNA sequencing were performed at days 0, 16, and 32. Results revealed an increased ∆BW in SC horses compared with CON horses (P = 0.03), with no change in BCS (P = 0.97). D-lactate tended to be greater in SC horses on day 32 at 16 and 24 h postmeal compared with CON horses (P = 0.10). Concentrations of TNFα and LogCCL2 decreased from day 0 to day 32 regardless of dietary supplementation (P ≤ 0.02). Fold change of percent reads from day 0 in whole blood bacterial 16S rRNA did not differ between groups. Fecal starch was undetectable, and there were no differences in intake or apparent digestibility. Fecal pH tended (P = 0.07) to be lower in CON at 0 h on day 32 (6.03 ± 0.06) than on day 16 (6.14 ± 0.06). Additionally, pH tended (P = 0.09) to be lower in CON (6.03 ± 0.06) than in SC (6.16 ± 0.06) at 0 h on day 32. Supplementation of S. cerevisiae CNCM I-1077 maintained Bacteroidales and reduced acidosis-like bacteria like Streptococcus and potential pathogens like Enterobacteriaceae, Stenotrophomonas, and Rhodococcus at day 16 (P < 0.05). Furthermore, supplementation increased fibrolytic bacteria at day 32, such as Ruminococcus, Fibrobacter, and Succinivibrio (P < 0.05). These results indicate S. cerevisiae CNCM I-1077 increases BW and promotes a more diverse microbiome when horses are fed ad libitum hay and a high-starch concentrate.

A Bioinformatics Tool for Identifying Intratumoral Microbes from the ORIEN Dataset

Evidence supports significant interactions among microbes, immune cells, and tumor cells in at least 10%-20% of human cancers, emphasizing the importance of further investigating these complex relationships. However, the implications and significance of tumor-related microbes remain largely unknown. Studies have demonstrated the critical roles of host microbes in cancer prevention and treatment responses. Understanding interactions between host microbes and cancer can drive cancer diagnosis and microbial therapeutics (bugs as drugs). Computational identification of cancer-specific microbes and their associations is still challenging due to the high dimensionality and high sparsity of intratumoral microbiome data, which requires large datasets containing sufficient event observations to identify relationships, and the interactions within microbial communities, the heterogeneity in microbial composition, and other confounding effects that can lead to spurious associations. To solve these issues, we present a bioinformatics tool, microbial graph attention (MEGA), to identify the microbes most strongly associated with 12 cancer types. We demonstrate its utility on a dataset from a consortium of nine cancer centers in the Oncology Research Information Exchange Network. This package has three unique features: species-sample relations are represented in a heterogeneous graph and learned by a graph attention network; it incorporates metabolic and phylogenetic information to reflect intricate relationships within microbial communities; and it provides multiple functionalities for association interpretations and visualizations. We analyzed 2,704 tumor RNA sequencing samples and MEGA interpreted the tissue-resident microbial signatures of each of 12 cancer types. MEGA can effectively identify cancer-associated microbial signatures and refine their interactions with tumors.

SIGNIFICANCE

Studying the tumor microbiome in high-throughput sequencing data is challenging because of the extremely sparse data matrices, heterogeneity, and high likelihood of contamination. We present a new deep learning tool, MEGA, to refine the organisms that interact with tumors.

Efficient computation of contributional diversity metrics from microbiome data with FuncDiv

MOTIVATION

Microbiome datasets with taxa linked to the functions (e.g. genes) they encode are becoming more common as metagenomics sequencing approaches improve. However, these data are challenging to analyze due to their complexity. Summary metrics, such as the alpha and beta diversity of taxa contributing to each function (i.e. contributional diversity), represent one approach to investigate these data, but currently there are no straightforward methods for doing so.

RESULTS

We addressed this gap by developing FuncDiv, which efficiently performs these computations. Contributional diversity metrics can provide novel insights that would be impossible to identify without jointly considering taxa and functions.

AVAILABILITY AND IMPLEMENTATION

FuncDiv is distributed under a GNU Affero General Public License v3.0 and is available at https://github.com/gavinmdouglas/FuncDiv.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.