Performance Analysis of de novo Metagenome Assemblers for Long-Reads HiFi Data

Contents

1. Overview
2. Metagenome Analysis Pipeline
   1. Quality Control
   2. Assembly
   3. Assessment
   4. Visualization
3. References

1. Overview

The following assemblers are benchmarked:

This project analyzes different de novo metagenome assemblers to assess their performance and limitations with respect to PacBio HiFi long-reads metagenome data (both synthetic and real).

The following assemblers are analyzed:

• metaFlye
• hiCanu
• hifiasm-meta
• Raven

2. Metagenome Analysis Pipeline

2.1 Quality Control

The assemblers were evaluated using the following public datasets:

	accession	#bases (Gb)	N50 read length (kb)	Median read QV	Sample description
E.coli	SRR10971019	1.4	...	...	E.coli K12
ATCC	SRR11606871	59.2	12.0	36	Mock, ATCC MSA-1003
zymoBIOMICS	SRR13128014	18.0	10.6	40	Mock, ZymoBIOMICS D6331
sheepA	SRR10963010	51.9	14.3	25	Sheep gut microbiome
humanV1	SRR15275211	18.8	11.0	39	Human gut, pool of 4 vegan samples
chicken	SRR15214153	33.6	17.6	30	Chicken gut microbiome

The sequencing data were uploaded to the Galaxy web platform¹, and the public server at usegalaxy.org was used to perform FastQC² analysis on the data. The results of the analysis have been publicly published here.

2.2 Assembly

Here, we will be running all 4 assemblers (metaFlye³, hiCanu⁴, hifiasm-meta⁵ and Raven⁶).

2.2.1 metaFlye Assembly

Command for running flye assembly:

./flye --pacbio-hifi /path/to/data/ecoli.fastq \
  --plasmids \
  --meta \ 
  --debug \ 
  --out-dir /path/to/output/ecoli_hifi/

Usage: flye 

--pacbio-hifi  for PacBio HiFi reads 
--plasmids   
--meta         for metagenome mode
--debuf        to enable debug output
--out-dir      output path

The shell script run_flye.sh would run the assembly tool on all the datasets.

2.2.2 hiCanu Assembly

Command for running canu assembly:

./canu \
  -p ecoli_hifi \
  -d path/to/output/ecoli_hifi/ \
  genomeSize=4.8m \ 
  -pacbio-hifi /path/to/data/ecoli.fastq 

Usage: canu 

-p            prefix for output
-d            output directory
genomeSize    approximate genome size to determine input reads coverage
-pacbio-hifi  for PacBio HiFi reads 

2.2.3 hifiasm-meta Assembly

Command for running hifiasm assembly:

./hifiasm_meta \
  -t10 \
  -o ecoli_hifi \
  /path/to/data/ecoli.fastq 2>ecoli_hifi.log

Usage : hifiasm

-t  set number of CPUs in use
-o  prefix of output files

2.2.4 Raven Assembly

Command for running raven assembly:

./raven \
  --threads=10 \
  /path/to/data/ecoli.fastq \
  --graphical-fragment-assembly ecoli.gfa > ecoli.fasta

Usage: raven
--graphical-fragment-assembly  prints the assembly graph in GFA format  
--threads                      number of threads  

2.3 Assessment

2.3.1 Assembly evaluation using metaQUAST

For assembly evaluation, we used metaQUAST⁷ to compute various metrics.

Command for running metaQUAST.

python3 metaquast.py \
    --threads=20 \
    --mgm \
    --gene-finding \
    --labels=metaFlye,hiCani,hifiasm-meta,Raven \
    /path/to/metaFlye/contigs ....
    -o /path/to/output

Usage: metaquast
--threads       maximum number of threads.
--mgm           force use of MetaGeneMark for gene finding.
--gene-finding  enables gene finding.
--labels        human-readable assembly names for the report.
-o              output directory. 

2.3.2 Binning using MetaBat2

We used MetaBat2⁸ for binning the assembly result to be further able to check for genome completedness.

Command for running MetaBat2:

./metabat2 \
  -i /path/to/data/ecoli_assembly.fasta \
  -o ecoli_hifi/bin \
  -d -v

Usage: metabat2
-i  Contigs in (gzipped) fasta file format 
-o  Base file name and path for each bin. The default output is fasta format.                   
-d  Debug output
-v  Verbose output

2.3.3 Genome Completeness using CheckM

In here, we will be evaluate the assemblies using CheckM⁹.

Some convention,

• Near-complete bin: $\geq 90%$ checkM completeness score & $\leq 5%$ contamination score.
• High-quality: $\geq 70%$ complete & $\leq 10%$ contaminated.
• Medium-quality: QS $> 50%$.

where QS (quality score) is given by completeness-(5*contamination).

Command for running CheckM:

./checkm lineage_wf \
  -t 40 \
  -x fa path/to/bins(MetaBat2) \
  path/to/output
  
./checkm analyze \
  -t 40 \
  checkm_data/hmms/phylo.hmm \
  -x fa \
  path/to/bins(MetaBat2) \
  path/to/output

./checkm qa \
  -t 40  \
  --out_format 1 \
  -f path/to/output/qa_result.txt \
  checkm_data/hmms/phylo.hmm \
  path/to/output

Usage: checkm
-t               number of threads
-x               input data extension
--output format  option 1
-f               file         

2.3.4 Quality Analysis using BUSCO

The quality of metagenome assembly was analyzed using BUSCO¹⁰.

Command for running BUSCO:

./busco \
  -f \
  -c 40 \
  -m genome \
  --augustus \
  -i path/to/input/assembly \
  -o path/to/output \
  --auto-lineage-euk
  
mkdir path/to/output/summary   
cp path/to/output/short_summary.* path/to/output/summary

python3 generate_plot.py \
  -wd path/to/output/summary

Usage: busco
-f                   force rewriting of existing files
-c                   specify the number of threads/cores to use
-m                   specify which BUSCO analysis mode to run (genome for genome assemblies)
--augustus           use augustus gene predictor for eukaryote runs
-i                   input sequence file in FASTA format
-o                   output directory
--auto-lineage-euk   run auto-placement just on eukaryote tree to find optimum lineage path
-wd                  location of your working directory

	>1Mb circular contigs	>1Mb circular contigs, near-complete	Near-complete MAGs	High-quality MAGs	Medium-quality MAGs
E.coli	x	x	x	x	x
ATCC	x	x	x	x	x
Zymo	x	x	x	x	x
Sheep	x	x	x	x	x
human	x	x	x	x	x
Chicken	x	x	x	x	x

	Wall clock (h)	PeakRSS (Gb)
E.coli	y	y
ATCC	y	y
Zymo	y	y
Sheep	y	y
Human	y	y
Chicken	y	y

2.4 Vizualization

For vizualization of complete circular contig from the, Bandage¹¹ was used.

Here's a Bandage plot of chicken's primary contig graph.

Assembler	Dataset	Node count	Edge count	Median depth
MetaFlye	E.coli	1	1	302.0x
	ATCC	137	68	192.0x
	Zymo	840	443	130.0x
	SheepA	27,296	14,735	10.0x
	HumanV1	24,671	17,056	9.0x
	Chicken	10,963	11,359	10.0x
HiCanu	E.coli	6,405	6,578	1.0x
	ATCC	82,270	95,069	1.0x
	Zymo	31,164	34,126	1.0x
	SheepA	505,391	499,733	1.0x
	HumanV1	370,372	360,885	1.0x
	Chicken	153,854	152,774	1.0x
Hifiasm-meta	E.coli	129	1	178.0x
	ATCC	72,765	126,045	11.0x
	Zymo	25,203	29,792	11.0x
	SheepA	34,783	5,624	3.0x
	HumanV1	23,114	3,720	3.0x
	Chicken	21,903	14,556	2.0x
Raven	E.coli	1	1	0.00134x
	ATCC	354	26	0.000612x
	Zymo	940	52	0.000488x
	SheepA	24,366	12,237	0.000349x
	HumanV1	15,703	2,676	0.000396x
	Chicken	8,635	1,721	0.000229x

3. References

Footnotes

1. Enis Afgan, Dannon Baker, Bérénice Batut, Marius van den Beek, Dave Bouvier, Martin Čech, John Chilton, Dave Clements, Nate Coraor, Björn Grüning, Aysam Guerler, Jennifer Hillman-Jackson, Vahid Jalili, Helena Rasche, Nicola Soranzo, Jeremy Goecks, James Taylor, Anton Nekrutenko, and Daniel Blankenberg. The Galaxy platform for accessible, reproducible and collaborative biomedical analyses: 2018 update, Nucleic Acids Research, Volume 46, Issue W1, 2 July 2018, Pages W537–W544, doi:10.1093/nar/gky379 ↩

2. Andrews, S. (n.d.). FastQC A Quality Control tool for High Throughput Sequence Data. http://www.bioinformatics.babraham.ac.uk/projects/fastqc/ ↩

3. Mikhail Kolmogorov, Derek M. Bickhart, Bahar Behsaz, Alexey Gurevich, Mikhail Rayko, Sung Bong Shin, Kristen Kuhn, Jeffrey Yuan, Evgeny Polevikov, Timothy P. L. Smith and Pavel A. Pevzner "metaFlye: scalable long-read metagenome assembly using repeat graphs", Nature Methods, 2020 doi:s41592-020-00971-x ↩

4. Nurk S, Walenz BP, Rhiea A, Vollger MR, Logsdon GA, Grothe R, Miga KH, Eichler EE, Phillippy AM, Koren S. HiCanu: accurate assembly of segmental duplications, satellites, and allelic variants from high-fidelity long reads. biorXiv. (2020) ↩

5. Feng, Xiaowen, et al. "Metagenome assembly of high-fidelity long reads with hifiasm-meta." arXiv preprint arXiv:2110.08457 (2021) ↩

6. Vaser, R., Šikić, M. Time- and memory-efficient genome assembly with Raven. Nat Comput Sci 1, 332–336 (2021). https://doi.org/10.1038/s43588-021-00073-4 ↩

7. Alla Mikheenko, Vladislav Saveliev, Alexey Gurevich, MetaQUAST: evaluation of metagenome assemblies, Bioinformatics, Volume 32, Issue 7, 1 April 2016, Pages 1088–1090, https://doi.org/10.1093/bioinformatics/btv697 ↩

8. Kang DD, Li F, Kirton E, Thomas A, Egan R, An H, Wang Z. 2019. MetaBAT 2: an adaptive binning algorithm for robust and efficient genome reconstruction from metagenome assemblies. PeerJ 7:e7359 https://doi.org/10.7717/peerj.7359 ↩

9. Parks DH, Imelfort M, Skennerton CT, Hugenholtz P, Tyson GW. 2015. CheckM: assessing the quality of microbial genomes recovered from isolates, single cells, and metagenomes. Genome Research, 25: 1043–1055. ↩

10. Mosè Manni, Matthew R Berkeley, Mathieu Seppey, Felipe A Simão, Evgeny M Zdobnov, BUSCO Update: Novel and Streamlined Workflows along with Broader and Deeper Phylogenetic Coverage for Scoring of Eukaryotic, Prokaryotic, and Viral Genomes, Molecular Biology and Evolution, Volume 38, Issue 10, October 2021, Pages 4647–4654. ↩

11. Wick R.R., Schultz M.B., Zobel J. & Holt K.E. (2015). Bandage: interactive visualisation of de novo genome assemblies. Bioinformatics, 31(20), 3350-3352. ↩