Microbiome differential abundance methods produce disturbingly different results across 38 datasets
Bibliographic record
Abstract
Abstract Identifying differentially abundant microbes is a common goal of microbiome studies. Multiple methods have been applied for this purpose, which are largely used interchangeably in the literature. Although it has been observed that these tools can produce different results, there have been very few large-scale comparisons to describe the scale and significance of these differences. In addition, it is challenging for microbiome researchers to know which differential abundance tools are appropriate for their study and how these tools compare to one another. Here, we have investigated these questions by analyzing 38 16S rRNA gene datasets with two sample groups for differential abundance testing. We tested for differences in amplicon sequence variants and operational taxonomic units (referred to as ASVs for simplicity) between these groups with 14 commonly used differential abundance tools. Our findings confirmed that these tools identified drastically different numbers and sets of significant ASVs, however, for many tools the number of features identified correlated with aspects of the tested study data, such as sample size, sequencing depth, and effect size of community differences. We also found that the ASVs identified by each method were dependent on whether the abundance tables were prevalence-filtered before testing. ALDEx2 and ANCOM produced the most consistent results across studies and agreed best with the intersect of results from different approaches. In contrast, several methods, such as LEfSe, limma voom, and edgeR, produced inconsistent results and in some cases were unable to control the false discovery rate. In addition to these observations, we were unable to find supporting evidence for a recent recommendation that limma voom, corncob, and DESeq2 are more reliable overall compared with other methods. Although ALDEx2 and ANCOM are two promising conservative methods, we argue that those researchers requiring more sensitive methods should use a consensus approach based on multiple differential abundance methods to help ensure robust biological interpretations.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.003 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".