4.2.2

Version 4.2.2 (Jun 4th, 2025)

Announcement: We are pleased to share that MetaPhlAn 4.2 is now available, which includes a new major database release, some code refactoring, and support for taxonomic profiling of long-read metagenomes produced by PacBio or nanopore platforms. This was made possible by integrating the Minimap2 aligner, which enables efficient mapping of long-read sequences. Improvements to the database include both the incremental inclusion of additional reference genomes (both isolates and MAGs) and the usage of an improved clustering method for more accurate binning of genomes into SGBs.

For details on MetaPhlan 4, check announcing MetaPhlAn 4 or visit the MetaPhlAn 4 GitHub repository.

For a complete inventory of the changes in MetaPhlAn 4.2, please see below:

Database updates

A new ChocoPhlAn database (“vJan25”) for MetaPhlAn and related tools has been released with MetaPhlAn 4.2. This new database includes:

• Addition of ~63k isolate genomes from NCBI, leading to a total of ~219k isolate genomes
• Addition of ~416k new MAGs, leading to a total of ~1,3M MAGs
• Expansion of the markers database with 58,331 SGBs (21,509 more than vJun23)
• Improved clustering of SGBs using skANI
• Addition of taxonomy assignment for Viral Sequence Clusters (VSCs) database performed with GeNomad
• Update SGB2GTDBtk file to r220 - allowing direct comparison between MetaPhlAn and GTDB results

New features

• [MetaPhlAn] Support for long-read sequencing datasets:
  Added “Long reads arguments” group of options:
  • --long_reads to profile a long-read dataset
  • --split_reads to profile long reads by splitting them and treating them as short reads
  • --split_readlen to specify the length of the splits when --split_reads is selected (default is 150)
  • --max_gcsd to set the threshold on gap-compressed sequence divergence to use when filtering the Minimap2 mapping results (default is 0.10)
  • --minimap2_exe to set a path to the Minimap2 executable
  • --minimap2_ps to change the mapping options for Minimap2 (defaults are '-x asm20 -B 3 -O 3,12')
  • --nbases required to specify the total number of bases in the sample when the input is a SAM file
• [MetaPhlAn] Added information about database used to run MetaPhlAn in SAM header as @ CO index:<database_index>
• [MetaPhlAn] Added implementation of RPKM to virus for short reads
• [MetaPhlAn] Added compatibility of --profile_vsc with --input_type sam
• [MetaPhlAn] Added --verbose option for debugging issues with database download
• [StrainPhlAn] Added in sample2markers.py an automatic detection of the mapper used for the alignment (Bowtie2 or Minimap2) to set the correct default values
• [StrainPhlAn] Added --max_gcsd option in sample2markers.py to set the threshold on gap-compressed sequence divergence to use when filtering markers from the Minimap2 mapping results (default is 0.10)
• [StrainPhlAn] Added --polymorphism_perc option in strainphlan.py to regulate the maximum percentage of polymorphic sites in a marker to be considered (default is None)

Changed features

• [MetaPhlAn] --unclassified_estimation is now a default parameter, making the relative abundance of detected taxa to be rescaled taking into account the estimated portion of unclassified reads. Use --skip_unclassified_estimation to not include an estimate of unclassified reads in the relative abundance profile
• [MetaPhlAn] Changed parameter for specifying the folder containing the MetaPhlAn database from --bowtie2db to --db_dir
• [MetaPhlAn] Changed parameter for specifying the path to the MetaPhlAn database folder from --bowtie2out to --mapout
• [MetaPhlAn] Changed value of --input_type from bowtie2out to mapout to specify the input is an intermediate mapping file from a previous run with --mapout (old --bowtie2out)
• [MetaPhlAn] Changed parameter for specifying output format from --biom to --biom_format_output. Now the parameter is a flag that allows the user to obtain the profiling in biom format only, instead of writing a separate output file. Also, --mdelim is now --biom_mdelim.
• [MetaPhlAn] Changed some default parameters for long reads:
  --min_mapq_val has default 5 for short reads and 50 for long reads
  --vsc_breadth has default 0.75 for short reads and 0.5 for long reads
  --subsampling takes as input the number of bases to be considered from the input metagenomes (instead of number of reads) if --long_reads is selected
• [MetaPhlAn] The MetaPhlAn species Newick trees for computing UniFrac distances in calculate_diversity.R are no longer stored in GitHub, but are downloaded with the database (or can be found here: http://cmprod1.cibio.unitn.it/biobakery4/metaphlan_databases/ )
• [StrainPhlAn] Changed some default parameters for long reads in sample2markers.py, depending on the mapper detected (Bowtie2 or Minimap2):
  --min_reads_aligning has default 8 for short reads and 1 for long reads
  --min_mapping_quality has default 10 for short reads and 50 for long reads
• Fixed bugs in 4.2.0 and 4.2.1

Dropped features

• [MetaPhlAn] Removed backward compatibility with --mpa3 which includes removal of --add_viruses
• [MetaPhlAn] Removed MetaPhlAn2 style output format --legacy-output

4.1.1

Version 4.1.1 (Mar 11th, 2024)

Database updates

• We just released the new vJun23_202403 and vOct22_202403 databases
  • Same SGBs as for the vJun23_202307 and vOct22_202212 versions, respectively, but the NCBI taxonomy assignment has been fixed to keep the taxa consistent across the MetaPhlAn taxonomic tree, allowing accurate relative abundance estimation also at higher taxonomic levels

New features

• [MetaPhlAn] The new fix_relab_mpa4.py script enables to fix errors in the relative abundances in profiles generated with previous databases
• [MetaPhlAn] Implementation of the option --subsampling_paired [N_PAIRED_READS] to subsample paired-end input reads. It needs to be used in conjunction with -1 [FORWARD_READS_FILE] and -2 [REVERSE_READS_FILE]

Fixes

• [MetaPhlAn] Fixed a bug that would halt MetaPhlAn execution when the option --profile_vsc was used but had no viral hits
• [MetaPhlAn] Fixed a bug that would halt MetaPhlAn execution when the number of reads to map was zero
• [StrainPhlAn] Fixed a bug in the new implementation (since v4.1) of –-print_clades_only

4.1.0

Version 4.1.0 (Feb 20th, 2024)

Database updates

• We just released the new vJun23 database
  • Addition of ~45k reference genomes from NCBI
  • Addition of ~50k MAGs from ocean, ~40k MAGs from soil, ~30k MAGs from domestic animals and non-human primates, ~4k MAGs from giant turtles, ~7.5k MAGs from skin microbiome, ~20k MAGs from dental plaque, ~15k MAGs from Asian populations, ~2.7k MAGs from ancient and modern Bolivians and other small datasets from diverse sources
  • Expansion of the markers database with 36,822 SGBs (6,272 more SGBs than in vOct22)
• Inclusion of the new Viral Sequence Clusters (VSCs) database
  • Containing 3,944 VSCs clustered into 1,345 Viral Sequence Groups (VSGs).
  • Including a total of 45,872 representative VSGs sequences.
  • Each cluster/group is labeled as known (kVSG) or unknown (uVSG) depending on the presence of at least a viral RefSeq reference genome within the cluster/group.

New features

• [MetaPhlAn] The new --profile_vsc parameter (together with --vsc_out and --vsc_breadth) enables the profiling of viral sequence clusters.
• [MetaPhlAn] The --subsampling now subsamples the FASTQ files and not the mapping results
• [MetaPhlAn] The new --mapping_subsampling parameter enables the previous mapping subsampling behaviour
• [MetaPhlAn] The new --subsampling_output parameter enables to save the subsampled FASTQ file
• [MetaPhlAn] The new create_toy_database.py script enables the custom filtering of the MetaPhlAn databases

Changed features

• [MetaPhlAn] The average read length is included in the output header with the -t rel_ab_w_read_stats parameter
• [StrainPhlAn] Quasi-markers behaviour in line with that of MetaPhlAn
• [StrainPhlAn] sample2markers.py output is now in JSON format
• [StrainPhlAn] Simplified sample and marker filtering parameters, integrated with primary/secondary samples
• [StrainPhlAn] Faster inference of small and medium phylogenies
• [StrainPhlAn] Faster execution of the parameter –-print_clades_only

4.0.6

Version 4.0.6 (Mar 1st, 2023)

Changed features

• [MetaPhlAn] The GTDB taxonomic assignment for the vOct22 database is now available.

4.0.5

Version 4.0.5 (Feb 23rd, 2023)

Database updates

• We just released the new vOct22 database
• Addition of ~200k new genomes
• 3,580 more SGBs than the vJan21
• 2,548 genomes considered reference genomes in vJan21 were relabelled as MAGs in NCBI -> 1,550 kSGBs in vJan21 are now uSGBs in vOct22
• Removed redundant reference genomes from the vJan21 genomic database using a MASH distance threshold at 0.1%
• Local reclustering to improve SGB definitions of oversized or too-close SGBs
• Improved GGB and FGB definitions by reclustering SGB centroids from scratch
• Improved phylum assignment of SGBs with no reference genomes at FGB level using MASH distances on amino acids to find the closest kSGB

Changed features

• [StrainPhlAn] Improved StrainPhlAn's speed when running with the --print_clades_only option

Missing features

• [MetaPhlAn] The GTDB taxonomic assignment for the vOct22 database is not available yet (expected release: end of Feb 2023)
• [MetaPhlAn] The phylogenetic tree of life for the vOct22 database is not available yet (expected release: TBD).

4.0.4

Version 4.0.4 (Jan 17nd, 2023)

Changed features

• [MetaPhlAn] Download of the pre-computed Bowtie2 database is now the default option during installation
• [StrainPhlAn] Improved StrainPhlAn's sample2makers.py script performance and speed

Fixes

• [StrainPhlAn] Fixes error when using --abs_n_samples_threshold in the PhyloPhlAn call

4.0.3

Version 4.0.3 (Oct 24nd, 2022)

Changed features

• [MetaPhlAn] Removal of the NCBI taxID from the merged profiles produced by the merge_metaphlan_profiles.py script
• [StrainPhlAn] Improved StrainPhlAn's performance in the markers/samples filtering step

Fixes

• [MetaPhlAn] -t rel_ab_w_read_stats now produces the reads stats also at the SGB level
• [MetaPhlAn] Fixes overstimation of reads aligned to known clades
• [MetaPhlAn] Fixes error when not providing the number of reads using SAM files as input
• [StrainPhlAn] Fixes No markers were found for the clade error while executing StrainPhlAn without providing the clade markers FASTA file

4.0.2

Version 4.0.2 (Sep 22nd, 2022)

New features

• [MetaPhlAn] The new --subsampling parameter allows reads' subsampling on the flight
• [MetaPhlAn] The new --subsampling_seed parameter enables a deterministic or randomized subsampling of the reads
• [MetaPhlAn] The new --gtdb_profiles of the merge_metaphlan_profiles.tsv allows the merge of GTDB-based MetaPhlAn profiles
• [StrainPhlAn] The new --breadth_thres parameter allows StrainPhlAn to filter the consensus markers sequences after the execution of sample2markers.py
• [StrainPhlAn] Interactive selection of the available SGBs when the clade is specified at the species level
• [StrainPhlAn] The new --non_interactive parameter disables user interaction when running StrainPhlAn
• [StrainPhlAn] The new --abs_n_markers_thres and --abs_n_samples_thres parameters enables the specification of the samples/markers filtering thresholds in absolute numbers
• [StrainPhlAn] The new --treeshrink parameter enables StrainPhlAn to run TreeShrink for outlier removal in the tree
• [StrainPhlAn] Addition of the VallesColomerM_2022_Jan21_thresholds.tsv for compatibility with the mpa_vJan21 database
• [StrainPhlAn] The new --clades parameter enables sample2markers.py to restrict the reconstruction of markers to the specified clades

Changed features

• [StrainPhlAn] The -c parameter of the extract_markers.py script now allows the specification of multiple clades
• [StrainPhlAn] The --print_clades_only parameter now produces an output print_clades_only.tsv report
• [StrainPhlAn] Compatibility with clade markers compressed in bz2 format
• [StrainPhlAn] The strain_transmission.py script now uses by the default the VallesColomerM_2022_Jan21_thresholds.tsv thresholds

Fixes

• [MetaPhlAn] metaphlan2krona.py and hclust2 have been added to the bioconda recipe

4.0.1

Changes in version 4.0.1

• The new --offline parameter stops MetaPhlAn from automatically checking for updates
• Fixes "KeyError: 't'" error when running MetaPhlAn with the --CAMI_format_output parameter
• Improved StrainPhlAn's gaps management with the newest version of PhyloPhlAn (version 3.0.3)
• Improved set of colors for the plot_tree_graphlan.py script

4.0.0

MetaPhlAn 4 relies on ~5.1M unique clade-specific marker genes identified from ~1M microbial genomes (~236,600 references and 771,500 metagenomic assembled genomes) spanning 26,970 species-level genome bins (SGBs), 4,992 of them taxonomically unidentified at the species level.

What's new in version 4

• Adoption of the species-level genome bins system (SGBs)
• New MetaPhlAn marker genes extracted identified from ~1M microbial genomes
• Ability to profile 21,978 known (kSGBs) and 4,992 unknown (uSGBs) microbial species
• Better representation of, not only the human gut microbiome but also many other animal and ecological environments
• Estimation of metagenome composed by microbes not included in the database with parameter --unclassified_estimation
• Compatibility with MetaPhlAn 3 databases with parameter --mpa3