MetaPhlAn /

Filename Size Date modified Message
893 B
biom Support: Conditionally importing modules required for biom output
5.7 KB
New supporting feature
1.1 KB
Licensing information added (MIT free software license, see license.txt)
2.6 MB
biom Support: Conditionally importing modules required for biom output
16.8 KB
New supporting feature
MetaPhlAn: Metagenomic Phylogenetic Analysis

Nicola Segata:

MetaPhlAn is a computational tool for profiling the composition of microbial communities from metagenomic shotgun sequencing data. 
MetaPhlAn relies on unique clade-specific marker genes identified from 3,000 reference genomes, allowing:
- orders of magnitude speedups compared to existing methods;
- unambiguous taxonomic assignments;
- accurate estimation of organismal relative abundance;
- species-level resolution for bacterial and archaeal organisms.

If you use this software, please cite our paper:
"Metagenomic microbial community profiling using unique clade-specific marker genes" 
Nicola Segata, Levi Waldron, Annalisa Ballarini, Vagheesh Narasimhan, Olivier Jousson, Curtis Huttenhower. 
Nature Methods, in press

* MetaPhlAn requires python 2.7 or higher with argparse, tempfile and numpy libraries installed 
  (apart for numpy they are usually installed together with the python distribution). 
  Python3 is not supported yet.

If you provide the output of BLASTN or BowTie2 as input, there are no additional prerequisite.

If you would like to use the integrated BLASTN in MetaPhlAn, you need to have:
* blastn from the NCBI BLAST+ package version 2.2.25 or higher (blastn needs to be in the system path)

If you would like to use the BowTie2 integrated in MetaPhlAn, you need to have:
* BowTie2 version 2.0.0 or higher and perl (bowtie2 needs to be in the system path with execute _and_ read permission)

If you use the "plotting_scripts/" script to plot and hierarchial cluster multiple metaphlan-profiled samples you will also need the following python libraries: matplotlib, scipy, pylab 


* Profiling a metagenome from raw reads using Blast (requires BLAST 2.2.25+ and the
 blast marker DB provided with MetaPhlAn): metagenome.fasta --blastdb blastdb/mpa

* You can save a lot of computational time if you perform the blasting using BowTie2 
  instead of Blast (requires BowTie2 in the system path with execution and read 
  permissions, Perl installed, and the BowTie2 marker DB provided with MetaPhlAn): metagenome.fasta --bowtie2db bowtie2db/mpa

* When possible, it is recommended to use fastq files that will increase the mapping 
  accuracy with BowTie2. No changes in the command line are required for using fastq 
  files, but a non-local hit policy for BowTie2 (i.e. '--bt2_ps sensitive-local') is 
  recommended for avoiding overly-sensitive hits: metagenome.fastq --bowtie2db bowtie2db/mpa --bt2_ps sensitive-local

* you can take advantage of multiple CPUs and you can save the blast output
   for re-running MetaPhlAn extremely quickly: metagenome.fasta --blastdb blastdb/mpa --nproc 5 --blastout metagenome.outfmt6.txt
  and the same for BowTie2: metagenome.fastq --bowtie2db bowtie2db/mpa --nproc 5 --bowtie2out metagenome.bt2out.txt

* if you already blasted your metagenome against the marker DB (using MetaPhlAn
   or blastn/BowTie2 alone) you can obtain the results in few seconds: --input_type blastout metagenome.outfmt6.txt
  (notice that 'blastout' define a file format independent from NCBI Blast and 
  common to BowTie2 as well)

* When using BowTie2 the metagenome can also be passed from the standard input but 
  it is necessary to specify the input format explicitly:
tar xjz metagenome.tar.bz2 --to-stdout | --input_type multifastq --blastdb blastdb/mpa

* Also the pre-computed blast/BowTie2 output can be provided with a pipe (again 
  specifying the input type): --input_type blastout < metagenome.outfmt6.txt > profiling_output.txt

* you can also set advanced options for the BowTie2 step selecting the preset option 
  among 'sensitive','very-sensitive','sensitive-local','very-sensitive-local' 
  (valid for metagenome as input only): --bt2_ps very-sensitive-local metagenome.fasta

* for for Blast the main configurable option is the threshold on the evalue
 (default 1e-5, we strongly recommend not lowering it too much): --evalue 1e-7 < metagenome.outfmt6.txt > profiling_output.txt

* if you suspect that the metagenome contains unknown clades, you may obtain
 more accurare result with a more sensible blast search lowering the blast
 word_size (the blasting will be slower): metagenome.fna --word_size 12 > profiling_output.txt


usage: [-h] [-v] [-t ANALYSIS TYPE] [--tax_lev TAXONOMIC_LEVEL]
                    [--nreads NUMBER_OF_READS] [--pres_th PRESENCE_THRESHOLD]
                    [--blastdb METAPHLAN_BLAST_DB]
                    [--bowtie2db METAPHLAN_BOWTIE2_DB] [--evalue]
                    [--word_size] [--bt2_ps BowTie2 presets] [--tmp_dir]
                    [--input_type {automatic,multifasta,multifastq,blastout}]
                    [--stat_q] [--blastn_exe BLASTN_EXE]
                    [--bowtie2_exe BOWTIE2_EXE] [--blastout FILE_NAME]
                    [--bowtie2out FILE_NAME] [--no_map] [-o output file]
                    [--nproc N]
                    [INPUT_FILE] [OUTPUT_FILE]

 MetaPhlAn version 1.7.7 (28 February 2013): METAgenomic PHyLogenetic ANalysis for
 taxonomic classification of metagenomic reads.

AUTHORS: Nicola Segata (

positional arguments:
  INPUT_FILE            the input file can be:
                        * a multi-fasta file containing metagenomic reads
                        * a NCBI BLAST output file (-outfmt 6 format) of the metagenome against the MetaPhlAn database. 
                        * a BowTie2 output file of the metagenome generated by a previous MetaPhlAn run 
                        The software will recognize the format automatically.
                        If the input file is missing, the script assumes that the input is provided using the standard 
                        input, and the input format has to be specified with --input_type
  OUTPUT_FILE           the tab-separated output file of the predicted taxon relative abundances 
                        [stdout if not present]

optional arguments:
  -h, --help            show this help message and exit
  -v, --version         Prints the current MetaPhlAn version and exit
  -t ANALYSIS TYPE      Type of analysis to perform: 
                         * rel_ab: profiling a metagenomes in terms of relative abundances
                         * reads_map: mapping from reads to clades (only reads hitting a marker)
                         * clade_profiles: normalized marker counts for clades with at least a non-null marker
                         * marker_ab_table: normalized marker counts (only when > 0.0 and normalized by metagenome size if --nreads is specified)
                         * marker_pres_table: list of markers present in the sample (threshold at 1.0 if not differently specified with --pres_th
                        [default 'rel_ab']
                        The taxonomic level for the relative abundance output:
                        'a' : all taxonomic levels
                        'k' : kingdoms (Bacteria and Archaea) only
                        'p' : phyla only
                        'c' : classes only
                        'o' : orders only
                        'f' : families only
                        'g' : genera only
                        's' : species only
                        [default 'a']
  --nreads NUMBER_OF_READS
                        The total number of reads in the original metagenome. It is used only when 
                        -t marker_table is specified for normalizing the length-normalized counts 
                        with the metagenome size as well. No normalization applied if --nreads is not 
                        Threshold for calling a marker present by the -t marker_pres_table option
                        The blast database file of the MetaPhlAn database 
  --bowtie2db METAPHLAN_BOWTIE2_DB
                        The BowTie2 database file of the MetaPhlAn database 
  --evalue              evalue threshold for the blasting
                        [default 1e-6]
  --word_size           word_size value for the blasting
                        [default NCBI BlastN default]
  --bt2_ps BowTie2 presets
                        presets options for BowTie2 (applied only when a multifasta file is provided)
                        The choices enabled in MetaPhlAn are:
                         * sensitive
                         * very-sensitive
                         * sensitive-local
                         * very-sensitive-local
                        [default very-sensitive-local]
  --tmp_dir             the folder used to store temporary files 
                        [default is the OS dependent tmp dir]
  --min_cu_len          minimum total nucleotide length for the markers in a clade for
                        estimating the abundance without considering sub-clade abundances
                        [default 10000]
  --input_type {automatic,multifasta,multifastq,blastouti,bowtie2out}
                        set wheter the input is the multifasta file of metagenomic reads or 
                        the blast output (outfmt 6 format) of the reads against the MetaPhlAn db.
                        [default 'automatic', i.e. the script will try to guess the input format]
  --stat_q              Quantile value for the robust average
                        [default 0.1]
  --blastn_exe BLASTN_EXE
                        Full path and name of the blastn executable. This option allows 
                        MetaPhlAn to reach the executable even when it is not in the system 
                        PATH or the system PATH is unreachable
  --bowtie2_exe BOWTIE2_EXE
                        Full path and name of the BowTie2 executable. This option allows 
                        MetaPhlAn to reach the executable even when it is not in the system 
                        PATH or the system PATH is unreachable
  --blastout FILE_NAME  The file for saving the output of the blasting (in outfmt 6 format)
  --bowtie2out FILE_NAME
                        The file for saving the output of BowTie2
  --no_map              Avoid storing the --blastout (or --bowtie2out) map file
  -o output file, --output_file output file
                        The output file (if not specified as positional argument)
  --nproc N             The number of CPUs to use for parallelizing the blasting
                        [default 1, i.e. no parallelism]


======== Version 1.7.7
New supporting feature
- the script conversion_scripts/ in conversion_scripts/ kindly implemented by Sami Pietila allows you to export MetaPhlAn output tables (merged with utils/ into the BIOM format. The BIOM format is the standard format for several post-processing programs and analyses such as Qiime.

======== Version 1.7.6
Bug fix
- now supporting --input_type bowtie2out that need to be specified as --input_type blastout before

======== Version 1.7.5
New supporting features:
- "-t marker_pres_table" outputs the list of markers with non-null abundance (i.e. present in the sample). The threshold for presence/absence call is specified via "--pres_th" 
- "-t marker_ab_table" outputs the normalize marker counts (multiplied by 1,000). The value is normalized by the size of the metagenome (to compare samples with different sequencing depth) if --nreads is specified

======== Version 1.7.4
New supporting feature:
- the script in plotting_scripts/ allows you to generate hierarchical clustering and heatmap visualization of multiple MetaPhlAn profiles. See plotting_scripts/readme.txt for details. 

======== Version 1.7.3
New feature:
- MetaPhlAn now accepts metagenomes from the standard input (when using BowTie2). When passing the input from the standard input, it is now required to specify the input type ('multifasta', 'multifastq', or 'blastout'). When using a file as input, the input type is still guessed by MetaPhlAn automatically.

======== Version 1.7.2
New supporting features
- multiple MetaPhlAn output obtained on distinct samples can now be easily merged into a unique "microbial clades vs samples" abundance table using the script in the "utils" folder (thanks to Tim Tickle)
- the MetaPhlAn output can now be visualized with Krona and with other software supporting the PhyloXML format, using the scritps in the "conversion_scripts" folder (thanks to Daniel Brami and James Taylor)

======== Version 1.7.1
Bug Fix
- two aspecific genes removed from the database (thanks to Nick Loman)

======== Version 1.7.0
New features
- added support for fastq files. Option available for analysis with BowTie2 only as blastn does not support fastq format. When possible, it is recommended to use fastq rather than fasta because BowTie2 will take into account the single nucleotide quality score contained in fastq format in the alignment evaluation. MetaPhlAn will automatically detect the fastq format so no changes in the command line options are required when using fastq files.

======== Version 1.6.0
New features
- added a Galaxy module that integrates MetaPhlAn into the galaxy environment. Source code and instructions for the module are provided, and the module currently runs at
- added --blastn_exe and --bowtie2_exe command line options for specifying the absolute path of the blastn and bowtie2 executables. This is useful when MetaPhlAn is used in a system without root priviledges and thus blastn and and bowtie2 have to be installed locally and not included into the system path.
- added the --no_map command line option for avoiding the blastn or bowtie2 output file to be saved on disk

======== Version 1.5.2
Bug fix:
- division by zero bug fixed: the bug was occurring when no hits were detected for any of the reads

======== Version 1.5.1
New options:
- option -o added for specifying the output file with a non-positional command line argument

======== Version 1.5
New features:
- added support for BowTie2 in addition to Blast. Bowtie2 runs at 10,000 reads-per-second (20,000 with 'sensitive' instead of 'very-sensitive-local') with accuracy performances at least comparable to blastn
-- '--bowtie2out' and '--bowtie2db' are the BowTie2 options corresponding to the '--blastout' and '--blastdb' blast options
-- support for the BowTie2 metaphlan output which is is simply a two-column file of reads/marker assignements. If called outside MetaPhlAn, BowTie2 should be called as 
   bowtie2 --quiet --sam-no-hd --sam-no-sq --very-sensitive-local --local -x bowtie2db/mpa -f -U metagenome.fna | cut -f 1,3 | grep -P -v "\*$"
-- the BowTie2 MetaPhlAn database in in bowtie2db, is called mpa, and consists of 6 files
-- four preset BowTie2 options are available (--bt2_ps option): --very-sensitive, --very-sensitive, --sensitive-local, --very-sensitive-local
- help message ' -h' expanded
- '--version' option added

Bug fixes:
- Fixed the error message for missing blastdb file (thanks to Joshua Reyes)

======== Version 1.1
New features and changes in MetaPhlAn version 1.1 with respect to version 1.0:
- no need to provide the MetaPhlAn data file anymore (thus --lib_dir has been removed)
- a robust average is introduced to more accurately estimate taxonomic relative abundances and dramatically lower false positives (--stat_q specifies the quantile for the truncated mean)
- the "word_size" option for blastn has been added in order to handle more accurately metagenomes without closely related reference genomes
- a new analysis type (-t clade_profiles) has been introduced in order to estimate normalized read counts for markers in each clade
- when using MetaPhlAn on blasting outputs, the input file can be provided using a pipe
- a more comprehensive help message has been added (see -h)
- the code has been re-factored and made more robust for future developments
- a slightly modified blastdb has been developed. As a consequence, the database (both multifasta and blastdb) of version 1.0 is not compatible with version 1.1