2019
DOI: 10.1101/847475
|View full text |Cite
Preprint
|
Sign up to set email alerts
|

Another look at microbe–metabolite interactions: how scale invariant correlations can outperform a neural network

Abstract: Many scientists are now interested in studying the correlative relationships between microbes and metabolites. However, these kinds of analyses are complicated by the compositional (i.e., relative) nature of the data. Recently, Morton et al. proposed a neural network architecture called mmvec to predict metabolite abundances from microbe presence. They introduce this method as a scale invariant solution to the integration of multi-omics compositional data, and claim that "mmvec is the only method robust to sca… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
4
1

Citation Types

0
12
0

Year Published

2019
2019
2022
2022

Publication Types

Select...
3
2

Relationship

0
5

Authors

Journals

citations
Cited by 7 publications
(12 citation statements)
references
References 10 publications
0
12
0
Order By: Relevance
“…As shown in the original MMvec paper [2] and Quinn and Erb et al [1], scale invariance is key for recovering sensible microbe-metabolite interactions. However, contrary to the scale invariance argument made in the preprint [1], MMvec does not normalize the joint distribution P (u i , v i ) between microbes and metabolites (microbe abundances are represented by u i and metabolite abundances are given by v i for sample i). Instead, the MMvec algorithm attempts to model P (v i |u i ) with an inverse alr transform, a known compositionally coherent transform that satisfies scale invariance [3].…”
Section: Responsementioning
confidence: 99%
See 4 more Smart Citations
“…As shown in the original MMvec paper [2] and Quinn and Erb et al [1], scale invariance is key for recovering sensible microbe-metabolite interactions. However, contrary to the scale invariance argument made in the preprint [1], MMvec does not normalize the joint distribution P (u i , v i ) between microbes and metabolites (microbe abundances are represented by u i and metabolite abundances are given by v i for sample i). Instead, the MMvec algorithm attempts to model P (v i |u i ) with an inverse alr transform, a known compositionally coherent transform that satisfies scale invariance [3].…”
Section: Responsementioning
confidence: 99%
“…Instead, the MMvec algorithm attempts to model P (v i |u i ) with an inverse alr transform, a known compositionally coherent transform that satisfies scale invariance [3]. This approach is more similar to a conditional version of approach B in the preprint rather than approach A [1]. Because of this, our method does not have the stated problem that microbe and metabolite abundances compete for probability mass in the normalized distribution: "the abundance of microbe 1 is limited by the abundance of microbes 2-to-M, but is in no way limited by the abundance of metabolites 1-to-N".…”
Section: Responsementioning
confidence: 99%
See 3 more Smart Citations