![Page 1: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/1.jpg)
Harvard School of Public Health
Department of Biostatistics
Meta’omic Analysis with MetaPhlAn & LEfSe
Eric Franzosa Postdoctoral Fellow / Huttenhower Lab
Symposium and Workshop on New Methods for Phylogenomics and Metagenomics
The University of Texas at Austin 17 February 2013
![Page 2: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/2.jpg)
Metagenomic Phylogenetic Analysis Fast and accurate metagenomic profiling of
microbial community composition using unique clade-specific marker genes
LDA Effect Size High-dimensional biomarker
discovery and explanation
http://http://huttenhower.sph.harvard.edu/content/metaphlan-tutorial
![Page 3: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/3.jpg)
3
Tutorial Outline
Introduction to MetaPhlAn
MetaPhlAn Demo
Introduction to LEfSe
LEfSe Demo
Links, other tools, and Q&A
![Page 4: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/4.jpg)
4
Tutorial Outline
Introduction to MetaPhlAn
MetaPhlAn Demo
Introduction to LEfSe
LEfSe Demo
Links, other tools, and Q&A
![Page 5: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/5.jpg)
Profile the taxonomic composition of microbial communities from whole shotgun metagenomic data
• Which clades (e.g. species, genera, classes) are there?
• What is the relative abundance of each clade in the community?
5
Tons of microbes
Shotgun sequencing
Tons of short reads
Taxonomic profiling
Relative abundances
?
MetaPhlAn Motivation
![Page 6: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/6.jpg)
Profile the taxonomic composition of microbial communities from whole shotgun metagenomic data
• Which clades (e.g. species, genera, classes) are there?
• What is the relative abundance of each clade in the community?
6
Several challenges
• Species-level resolution
• Computationally feasibility
• Organismal relative abundance rather than DNA concentrations
• Consistent detection confidence for all clades, including archaea
• High accuracies for very short reads (as short as ~50nt)
• Detection of organisms without sequenced genomes at higher taxonomic levels
Tons of microbes
Shotgun sequencing
Tons of short reads
Taxonomic profiling
Relative abundances
?
MetaPhlAn Motivation
![Page 7: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/7.jpg)
MetaPhlAn Overview
A X Y X B Y X Y C
Genomes
Sho
rt R
ead
s
7
![Page 8: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/8.jpg)
MetaPhlAn Overview
A X Y X B Y X Y C
Genomes
Sho
rt R
ead
s
8
![Page 9: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/9.jpg)
MetaPhlAn Overview
A X X B X Y C
Genomes
Sho
rt R
ead
s
A Y X B Y Y C
9
![Page 10: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/10.jpg)
10
Gene A
A is a core gene for clade Y A is a unique marker gene for clade Y
MetaPhlAn Overview
![Page 11: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/11.jpg)
11
Gene A
A is a core gene for clade Y A is a unique marker gene for clade Y
MetaPhlAn Overview
![Page 12: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/12.jpg)
12
Tutorial Outline
Introduction to MetaPhlAn
MetaPhlAn Demo
Introduction to LEfSe
LEfSe Demo
Links, other tools, and Q&A
![Page 13: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/13.jpg)
13
MetaPhlAn Demo: Setup
Download program and marker database from: http://huttenhower.sph.harvard.edu/metaphlan
Requires python with numpy installed
Requires BLAST or Bowtie2 for alignment
(Bowtie2 recommended)
Today’s sample data available at: http://huttenhower.sph.harvard.edu/content/metaphlan-tutorial
![Page 14: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/14.jpg)
14
MetaPhlAn (command line)
Show minimal MetaPhlAn setup
Run MetaPhlAn on downsampled HMP FASTA
Examine marker output file
Example abundance output file
Discuss taxonomic organization
Discuss “unclassified” species
![Page 15: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/15.jpg)
15
MetaPhlAn (command line)
![Page 16: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/16.jpg)
16
MetaPhlAn (command line)
![Page 17: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/17.jpg)
17
MetaPhlAn (command line)
![Page 18: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/18.jpg)
18
MetaPhlAn (command line)
![Page 19: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/19.jpg)
19
MetaPhlAn (command line)
![Page 20: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/20.jpg)
20
Tutorial Outline
Introduction to MetaPhlAn
MetaPhlAn Demo
Introduction to LEfSe
LEfSe Demo
Links, other tools, and Q&A
![Page 21: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/21.jpg)
21
Turning anecdotes into biology
![Page 22: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/22.jpg)
22
Turning anecdotes into biology
• More samples are a start… • But the statistics are still non-trivial
• Data are noisy • Compositional nature (Σ = 1) • High dynamic range • Hierarchical organization
![Page 23: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/23.jpg)
23
Turning anecdotes into biology
• We may also want to leverage (or control for) additional metadata
![Page 24: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/24.jpg)
24
Statistical consistency
Biological consistency
Biological Effect size
LEfSe: finding metagenomic biomarkers
![Page 25: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/25.jpg)
25
Tutorial Outline
Introduction to MetaPhlAn
MetaPhlAn Demo
Introduction to LEfSe
LEfSe Demo
Links, other tools, and Q&A
![Page 26: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/26.jpg)
26
LEfSe Demo: Galaxy Version
Available at: http://huttenhower.sph.harvard.edu/galaxy/
![Page 27: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/27.jpg)
27
Load your data table
Tab delimited text, consisting of a class (e.g. oral vs. gut), an optional subclass (e.g. male vs. female), an optional sample ID, and then your features.
![Page 28: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/28.jpg)
28
My input to Galaxy…
• MetaPhlAn output for many HMP subjects • Buccal_mucosa is a proxy for oral • Stool is a proxy for gut • Input file included with the sample data
![Page 29: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/29.jpg)
29
My input to Galaxy…
• MetaPhlAn output for many HMP subjects • Buccal_mucosa is a proxy for oral • Stool is a proxy for gut • Input file included with the sample data
• Features don’t have to come from MetaPhlAn • Any tab-delimited continuous data will work!
![Page 30: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/30.jpg)
30
Tell LEfSe which rows are class/subclass/ID
We’re just using class (STSite, Stool versus Buccal_mucosa) Set the other boxes to “no subclass” and “no subject”
![Page 31: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/31.jpg)
31
Run the algorithm
Because the oral and gut communities are quite different, I adjusted the threshold from the default of 2 to 4.5 (logs) to only
show the most extreme differences.
![Page 32: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/32.jpg)
32
Visualize the results
![Page 33: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/33.jpg)
33
LEfSe Output: Cladogram view
![Page 34: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/34.jpg)
34
LEfSe Output: Raw Numbers
![Page 35: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/35.jpg)
35
LEfSe Output: feature-specific report
![Page 36: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/36.jpg)
36
LEfSe at the command line
Source code available at: https://bitbucket.org/nsegata/lefse
Requires python and R with several additional
packages installed (see README)
Today’s sample data available at: http://huttenhower.sph.harvard.edu/content/metaphlan-tutorial
We’ll look at a second example included with the LEfSe distribution involving subclasses
![Page 37: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/37.jpg)
37
LEfSe (command line) demo
Show LEfSe setup
Show example folder
Highlight LEfSe commands
Show output
Highlight more complicated subclass treatment
![Page 38: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/38.jpg)
38
LEfSe (command line) demo
![Page 39: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/39.jpg)
39
Tutorial Outline
Introduction to MetaPhlAn
MetaPhlAn Demo
Introduction to LEfSe
LEfSe Demo
Links, other tools, and Q&A
![Page 40: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/40.jpg)
40
If you’re thirsty for more…
Tutorial materials available here: http://huttenhower.sph.harvard.edu/content/metaphlan-tutorial
My email (Eric Franzosa): [email protected]
![Page 41: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/41.jpg)
41
If you’re thirsty for more…
For more about our tools, check out: http://huttenhower.sph.harvard.edu/research
Individual tool pages contain links to source code and documentation…
…or try them out on Galaxy: http://huttenhower.sph.harvard.edu/galaxy/
![Page 42: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/42.jpg)
42
If you’re thirsty for more…
MetaPhlAn 2.0 coming soon with support for viruses, eukaryotes, and more bacteria/archaea
PhyloPhlAn uses MetaPhlAn’s core gene identification pipeline to assist in tree building http://huttenhower.sph.harvard.edu/phylophlan
MaAsLin is an evolution of LEfSe that can consider
a wider number and variety of metadata http://huttenhower.sph.harvard.edu/maaslin
![Page 43: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/43.jpg)
Ramnik Xavier
Harry Sokol
Dan Knights
Moran Yassour
43
Nicola Segata
Levi Waldron
Human Microbiome Project Owen White
Joe Petrosino
George Weinstock
Karen Nelson
Lita Proctor Dirk Gevers
Kat Huang
Bruce Birren Mark Daly
Doyle Ward Ashlee Earl
http://huttenhower.sph.harvard.edu
Joseph Moon
Vagheesh
Narasimhan
Tim Tickle Xochi Morgan
Jacques Izard
Katherine Lemon
Wendy Garrett
Michelle Rooks
Daniela
Boernigen
Bruce Sands
Rob Knight
Greg Caporaso
Jesse Zaneveld
Mark Silverberg
Boyko Kabakchiev
Andrea Tyler
Ruth Ley
Omry Koren
Emma
Schwager
Jim Kaminski
Brian Palmer
Craig Bielski
Ren Lu Hufeng Zhou
Sahar Abubucker
Brandi Cantarel
Alyx Schubert
Mathangi Thiagarajan
Beltran Rodriguez-Mueller
Erica Sodergren
Anthony Fodor
Marty Blaser
Jacques Ravel
Pat Schloss
Makedonka Mitreva
Yuzhen Ye
Mihai Pop
Larry Forney
Barbara Methe
Rob Beiko
Morgan Langille
Thank you!
Curtis
Huttenhower
![Page 44: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/44.jpg)
44
If you’re thirsty for more…
Questions?
![Page 45: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/45.jpg)
45
• Representing 2,887 genomes (107 Archaea)
• 1,222 species, 652 genera, 278 families, 130 orders, 66 classed, 33 phyla, 2,383 total clades
• ≈2M total unique marker genes
• ≈400k “most representative” unique marker genes
• 231±107 markers per species (350 fixed max)
• The 400k database represent ≈4% of the total microbial sequence data available
MetaPhlAn Statistics
![Page 46: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/46.jpg)
46
Evaluation of accuracy (1)
Species Classes
(Validation on high-complexity uniformly distributed synthetic metagenomes.)
![Page 47: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/47.jpg)
47
Evaluation of accuracy (2)
Species Classes
(Validation on low-complexity log-normally distributed synthetic metagenomes.)
![Page 48: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/48.jpg)
48
MetaPhlAn rel. abundance
16
S re
l. ab
un
dan
ce
Evaluation of accuracy (3)
• Buccal mucosa example
• Genus-level comparison with 16S-based estimation
• v13 ( + ) & v35 (×) regions
![Page 49: Meta’omic Analysis with MetaPhlAn & LEfSeHarvard School of Public Health Department of Biostatistics Meta’omic Analysis with MetaPhlAn & LEfSe Eric Franzosa Postdoctoral Fellow](https://reader033.vdocuments.us/reader033/viewer/2022041511/5e27fd7f89cee426ed5a2d63/html5/thumbnails/49.jpg)
49
>50 times faster than existing methods
450 reads/sec (BLAST)
Up to 25,000 reads/sec (bowtie2)
Multi-threaded
Easily parallelizable
Evaluation of performance