MetaPhlAn4

What does it do?

Use MetaPhlAn4 to profile the taxonomic composition of microbial communities from metagenomic shotgun sequence data.

The automator supports:

  • draft assemblies under fastas;
  • short-read FASTQ data under fastqs;
  • Nanopore FASTQ data under nanoporefastqs.

A single request can include more than one input type. If both an assembly and its raw reads must be analyzed, list the same SEQID under each applicable input-type heading.

For background, see the MetaPhlAn website, the MetaPhlAn 4 tutorial, and Blanco-Míguez et al. (2023). Cite the MetaPhlAn authors when publishing results produced with this automator.

How do I use it?

Subject

In the Subject field, enter:

metaphlan

Spelling matters, but matching is not case-sensitive.

Description

Group each SEQID under one of these input-type headings:

fastqs
fastas
nanoporefastqs

Example containing short reads and an assembly:

fastqs
2023-SEQ-0415
2023-SEQ-0414
fastas
2025-MIN-0333

To analyze both the assembly and raw data for the same sample, repeat its SEQID under both applicable headings.

Attachments

No attachment is required. The automator retrieves the requested assemblies or read files according to the input-type headings.

Optional parameters

analysis

Selects the MetaPhlAn output mode.

  • Default: rel_ab_w_read_stats
  • Supported values:
  • analysis=rel_ab_w_read_stats — relative abundances plus estimated reads from each clade;
  • analysis=rel_ab — relative-abundance profile;
  • analysis=clade_profiles — normalized marker counts for clades with at least one non-null marker;
  • analysis=marker_ab_table — normalized marker counts;
  • analysis=marker_pres_table — list of markers present in the sample.

Examples

Relative abundance with read statistics

analysis=rel_ab_w_read_stats
fastqs
2026-SEQ-0001
2026-SEQ-0002

Compare raw reads and an assembly for one sample

analysis=rel_ab
fastqs
2026-SEQ-0001
fastas
2026-SEQ-0001

Nanopore reads

nanoporefastqs
2026-MIN-0001

See issue 37794 for an example MetaPhlAn4 request. Temporary Dropbox links associated with the issue may expire.

Interpreting results

When MetaPhlAn4 finishes, it uploads an archive named using the Redmine issue identifier:

metaphlan4_output_redmineID.zip

The archive contains MetaPhlAn4 report files. The source documentation also states that it contains Bracken analysis reports; because Bracken is normally associated with Kraken classification, confirm the current archive contents before final publication.

Interpret the output according to the selected analysis mode:

  • rel_ab_w_read_stats reports relative abundance and estimated read counts per clade;
  • rel_ab reports relative abundance;
  • clade_profiles reports normalized marker counts for detected clades;
  • marker_ab_table reports normalized marker abundance;
  • marker_pres_table reports marker presence.

Results can differ between raw reads and assemblies because assembly may discard, combine, or fail to reconstruct some community sequences. Treat repeated analysis of one SEQID under different input headings as separate evidence rather than interchangeable output.

How long does it take?

Runtime depends on input type, selected analysis mode, amount of sequence data, and number of requested samples.

What can go wrong?

A requested SEQID is unavailable

Symptom: The Redmine issue receives a warning identifying unavailable sequences.

Likely cause: The automator cannot locate data matching the input-type heading used for the SEQID.

What to do: Verify the identifier and place it under fastqs, fastas, or nanoporefastqs according to the available data.

A SEQID is listed under the wrong input type

Symptom: The automator cannot retrieve or process the requested sequence data.

Likely cause: An assembly was listed under a raw-read heading, or raw reads were listed under fastas.

What to do: Move the SEQID beneath the correct input-type heading and resubmit the request.

Only one representation of a sample is analyzed

Symptom: The output contains raw-read results but no assembly result, or the reverse.

Likely cause: The SEQID was listed under only one input-type heading.

What to do: List the same SEQID under every representation that must be analyzed.

An unsupported analysis mode is requested

Symptom: The automator rejects the analysis or produces no expected report.

Likely cause: The analysis value is misspelled or unsupported.

What to do: Use one of the documented analysis values exactly as shown.

  • Kraken2/Bracken — provides k-mer-based taxonomic classification and refined abundance estimation with selectable databases.
  • AutoCLARK — reports species represented in raw reads or draft assemblies using CLARK.
  • StrainMash — identifies the closest RefSeq type strain for an assembly rather than profiling a mixed community.