Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Editorial
. 2024 Jun 17:4:1432791.
doi: 10.3389/fsysb.2024.1432791. eCollection 2024.

Specialty grand challenge: how can we use integrative approaches to understand microbial community dynamics?

Umer Zeeshan Ijaz  1   2   3 , Aqsa Ameer  1   4 , Farrukh Saleem  4   5 , Farzana Gul  4 , Ciara Keating  6 , Sundus Javed  4
Affiliations
Editorial

Specialty grand challenge: how can we use integrative approaches to understand microbial community dynamics?

Umer Zeeshan Ijaz et al. Front Syst Biol. .
No abstract available

Keywords: ecology; integrative “omics”; microbiology; multivariate statistical analyses; regression analysis.

PubMed Disclaimer

Conflict of interest statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The author(s) declared that they were an editorial board member of Frontiers, at the time of submission. This had no impact on the peer review process and the final decision.

Figures

FIGURE 1
FIGURE 1
The CODA-LASSO approach extends the linear regression approach by taking log abundances of microbiome data and considering two additional constraints. Constraint 1 is the LASSO constraint that forces some of the beta coefficients to go to zero allowing for variable selection, while constraint 2 ensures that there are two subsets of beta coefficients of microbes, those that are positively associated with the outcome, and those that are negatively associated with the outcome.
FIGURE 2
FIGURE 2
GLLVM procedure that fits a distribution for each microbe and regresses against all covariates to obtain beta coefficients that reveal positive or negative relationships.
FIGURE 3
FIGURE 3
Mediation analysis where the path from Treatment to Outcome is established after solving the inverse regression problem and testing the corresponding beta coefficients that exist on that path.
FIGURE 4
FIGURE 4
BIOENV approach, where the similarity distance is calculated for a fixed matrix (microbiome) and the features in the variable table are permuted to calculate a variable similarity distance in such a way that those subsets are retained where the correlation between the distances of the two matrices is optimized.
FIGURE 5
FIGURE 5
Constrained ordinations, where the left side visualization is obtained for CCA/RDA with the length of the arrows pointing to the direction of increase of the continuous covariates, while the right side is the smooth contour of a continuous covariate on the reduced ordination space. It should be noted that for CCA/RDA categorical variables can also be used where we get separate vectors for each factor of a categorical variable.
FIGURE 6
FIGURE 6
The Ensemble Quotient optimization algorithm works by finding an ensemble (a subset of microbes) that either remains stable (Uniform), has a stratified response (Categorical), or correlates with a covariate of interest (Continuous).
FIGURE 7
FIGURE 7
(A) Given pi , the proportional abundance of a microbe in the i -th environment, ri an environmental property (continuous covariate) is associated with each environment. (B) In the total number of environments being R (stratification based on some commonality among samples), we can assume for pi , and ri to be fairly stable within the environment and vary more across the environments. (C) Hurlbert’s BN and Feinsinger’s PS statistic indicates a negative or positive relationship between a microbe and an environmental property. A null modeling procedure is considered by generating a random normal distribution with 999 possible estimates, and by labeling a microbe as “negative” if the metric <fifth Quantile, and “positive”, if the metric >95th Quantile. (D) For Proportional Overlap POi,j , X (for microbe i) and Y (for microbe j) are Feinsinger’s PS.
FIGURE 8
FIGURE 8
(A) Traditional way of finding an association (whether negative or positive) between microbial abundance and a continuous covariate of interest. (B) The goal of specificity analysis is to find microbes that become specific within a narrow range of a covariate of interest. (C) Creating a null distribution utilizing Rao’s Quadratic Entropy formula (which incorporates the difference of the covariate across two samples i and j as Dij , multiplied by the relative abundance of the microbe across two samples as pi and pj ). These are calculated for all pairwise combinations of samples ( ss−12 ). The empirical value is then compared with the average of the 999 null distribution values to tag a microbe as specific (Spec ∼ −1) or cosmopolitan (Spec ∼ +1).
FIGURE 9
FIGURE 9
The core microbiome is established by first ranking the microbes based on a combination of two criteria, Site-specific Occupancy and Replicate Consistency, and then using the Bray-Curtis (BC) contribution of the core subset to the whole community; the core subset is iteratively constructed until the stopping criteria are reached. The core subset is then further discretized to those that: fall above the neutral model (selected by host/environment and represented by red dots in the bottom right figure); fall below the neutral model (selected by dispersal limitation and represented by blue dots); and those that fit the 95% CI of the neutral model (represented by green dots).
FIGURE 10
FIGURE 10
Illustration of the sPLSDA algorithm and its extension. The optimization algorithms are shown for all cases along with the weighting strategy for DIABLO shown in a table.
See this image and copyright information in PMC

Similar articles

See all similar articles

References

    1. Ainsworth T. D., Krause L., Bridge T., Torda G., Raina J. B., Zakrzewski M., et al. (2015). The coral core microbiome identifies rare bacterial taxa as ubiquitous endosymbionts. ISME J. 9 (10), 2261–2274. 10.1038/ISMEJ.2015.39 - DOI - PMC - PubMed
    1. Alekseyenko A. V. (2016). Multivariate Welch t-test on distances. Bioinforma. Oxf. Engl. 32 (23), 3552–3558. 10.1093/BIOINFORMATICS/BTW524 - DOI - PMC - PubMed
    1. Anderson M. J., Legendre P. (1999). An empirical comparison of permutation methods for tests of partial regression coefficients in a linear model. J. Stat. Comput. Simul. 62 (3), 271–303. 10.1080/00949659908811936 - DOI
    1. Andries E., Nikzad-Langerodi R. (2022). Dual-constrained and primal-constrained principal component analysis. J. Chemom. 36 (5), e3403. 10.1002/CEM.3403 - DOI
    1. Armengaud J. (2023). Metaproteomics to understand how microbiota function: the crystal ball predicts a promising future. Environ. Microbiol. 25 (1), 115–125. 10.1111/1462-2920.16238 - DOI - PMC - PubMed

Publication types

LinkOut - more resources