Skip to content
This repository has been archived by the owner on Jun 10, 2024. It is now read-only.
/ SSAKE Public archive

๐Ÿถ Genome assembly with short sequence reads

License

Notifications You must be signed in to change notification settings

bcgsc/SSAKE

Folders and files

NameName
Last commit message
Last commit date

Latest commit

ย 

History

79 Commits
ย 
ย 
ย 
ย 
ย 
ย 
ย 
ย 

Repository files navigation

Release Downloads Conda Issues link Thank you for your Stars

Logo

Short Sequence Assembly by K-mer search and 3' read Extension

Rene L Warren

2006-present

Contents


  1. Description
  2. For best performance
  3. What's new
  4. Implementation and requirements
  5. Installation
  6. Documentation
  7. Citing SSAKE
  8. Running SSAKE
  9. Test data
  10. Algorithm
  11. Input
  12. Output
  13. License

Description


SSAKE is a genomics application for de novo assembly of millions of very short DNA sequences.

Algorithms of SSAKE are the core of early genomics applications that followed (e.g., VCAKE, QSRA, SHARCGS, SSPACE, JR-Assembler) and their design have endured in other work (e.g., JR-Assembler, PNAS 2013; ViQuaS, Bioinformatics 2015; novoBreak, Nature Methods 2017). Applications of SSAKE extend beyond genome assembly and the technology was applied to profiling T-cell metagenomes, targeted de novo assembly (TASR), proteome assembly (PASS), genome scaffolding (LINKS, ARCS), HLA typing (HLAminer) and was key to the discovery of Fusobacterium in colon cancer, a finding designated as one of the top 10 medical breakthroughs of 2011 by Time magazine. SSAKE is also the assembly engine in the top-performing cancer genomic structural variant predictor pipeline software novoBreak, as assessed by the ICGC-TCGA DREAM Genomic Mutation Calling Challenge.

For best performance


Best assembly results are achieved with quality-trimmed reads. When working with Solexa/Illumina sequences, remove low quality bases, whenever possible, with:

~/ssake_v4.0/tools/
TQS.py 
eg. TQS.py -f _seq.txt -q _prb.txt -t 5 -d 5 -l #CYCLES -c 20  OR OTHER SETTINGS 
TQSfastq.py
eg. TQSfastq.py -f myfile.fq -t 20 -c 30 -e 64
TQSfastq.pl
eg. TQSfastq.pl -f human.fof -q 20 -n 70 -e 33
where human.fof contains fastq files to trim
make sure the trimmed fastq ends with 1 and 2 for the first and second reads of a pair
eg. makePairedOutput2UNEQUALfiles.pl NA24143_genome_phased_namesorted.bam1_1.fqc70q20e33.fa NA24143_genome_phased_namesorted.bam1_2.fqc70q20e33.fa 350
will produced "paired.fa" and "unpaired.fa"
~/ssake_v4.0/SSAKE -f paired.fa -g unpaired.fa -p 1 -m 20 -o 3 -c 1 -w 5

A BASIC PIPELINE TO HELP YOU PREPARE YOUR INPUT PAIRED READS EXISTS (./tools/runSSAKE.sh)

USAGE: ./runSSAKE.sh read1.fq read2.fq libraryFragmentLength basename

The scripts are located in ./tools subdirectory included with this release. It is recommended that you run TQS.py/TQSfastq.py for every tile (batch job) and cat the outputted fasta file, especially if your data set is large (e.g. entire flowcell)

For trimming paired-end sequences (using _seq.txt and _prb.txt from Illumina), please refer to TRIMMING_PAIRED_READS.README located in the ./tools subdirectory

What's new in v4.0 ?


-Initial support for linked reads, such as those produced by the 10x Genomics Chromium technology
-Linked reads are supplied in the original SSAKE format, but the barcode information is indicated by sequence following a single underscore "_"
-Added a C. elegans linked read data assembly script in the test folder
-Scaffolding options consistent with that of the LINKS scaffolder
-SSAKE no longer sorts by/prioritizes high multiplicity reads for assembly
-Streamlined file names
-Deprecated force-fill gap feature
-Bug fixes

What's new in v3.8.5 ?


Implements targeted de novo assembly. Fixed a bug that manifested when running SSAKE in targeted (-s) de novo assembly mode. Contig sequences as long as the seed sequence were previously misclassified as singlets.

What's new in v3.8.4 ?


Improvements to the targeted assembly functionality, recruiting whole read pairs for de novo assembly when at least one read has a k-mer match. This has the potential to extend the reconstructed contigs by 2X the library fragment size (upstream and downstream) of the target sequence when run in targeted de novo assembly mode -s and -i 1 (default).

What's new in v3.8.3 ?


Included tie-breaker option (-q) when determining consensus from equal-coverage bases. When set (-q 1), the parameter r is overridden. Note that even though the base is "randomly" chosen in such situations, SSAKE will only extend contigs in cases where there is read support over the chosen base. Included option to ignore reads making up the consensus base extension (-y).

What's new in v3.8.2 ?


Included target word length (-j option) - TASR behaviour (equivalent to -k option in TASR)

What's new in v3.8.1 ?


fixed SSAKE for Perl >= 5.16.0, where deprecated getopts.pl has been removed. Thanks to Nicola Soranzo for sending the fix.

What's new in v3.8+ ?


This release is 30% faster than v3.7, made possible by 4-base encoding of the first 16 bases of each read while populating a 4-nodes (16/4) prefix tree.  Also, the newly implemented (and required) -w option, gives users more control over the assembly, focusing on higher-depth contigs and ignoring short, low-depth contigs comprised of NGS reads having errors, contaminating reads or any other (perhaps unwanted) sequences. 

*The assembly of 25M high-quality (100bp x 2 x Q30) bacterial NGS reads (1 lane, Illumina HiSeq) with SSAKE v3.7 in single-end mode took 2h50.  SSAKE v3.8 ran for 2h on the same data.  A SSAKE v3.8 assembly restricted to high-depth contigs (-w 10) ran in 33 minutes, and eliminated thousands of short, low depth of coverage contigs that would otherwise arise from contaminating reads and reads with base errors.    

*The assembly of ~1M quality-trimmed (../tools/TQSfastq.py -f Assemble_1_R1.fastq -t 30 -c 100 -e 33) Campylobacter showae CC57C (BioProject/Accession:PRJNA189774/AOTD00000000) bacterial NGS reads (1 lane, Illumina MiSeq, PE151, 1.8M pairs sequenced) with SSAKE v3.8.2 in paired-end mode took 10m31s and 3.8GB RAM on a 12-core 48GB RAM machine (Benchmark with syrupy @https://github.com/jeetsukumaran/Syrupy) and yielded 215 contigs with N50=41kbp (151 scaffolds, N50=124kbp), and a reconstruction of 2.2Mbp. 

Campylobacter showae CC57C colorectal cancer tumor isolate (~2Mbp genome) Illumina MiSeq TEST DATA AVAILABLE HERE:
http://www.bcgsc.ca/downloads/supplementary/SSAKE/CC57C_paired.fa and CC57C_unpaired.fa
 
SSAKE ASSEMBLY PIPELINE:
./test/runSSAKE.sh

./tools/TQSfastq.py -f Assemble_1_R1.fastq -t 30 -c 100 -e 33
./tools/TQSfastq.py -f Assemble_1_R2.fastq -t 30 -c 100 -e 33
cat Assemble_1_R2.fastq_T30C100E33.trim.fa |perl -ne 'if(/^(\>\@\S+)/){print "$1b\n";}else{print;}' >Assemble_1_R2.fastq_T30C100E33.trimFIX.fa
cat Assemble_1_R1.fastq_T30C100E33.trim.fa |perl -ne 'if(/^(\>\@\S+)/){print "$1a\n";}else{print;}' >Assemble_1_R1.fastq_T30C100E33.trimFIX.fa
./tools/makePairedOutput2UNEQUALfiles.pl Assemble_1_R1.fastq_T30C100E33.trimFIX.fa Assemble_1_R2.fastq_T30C100E33.trimFIX.fa 400
./Syrupy-1.4.0/scripts/syrupy.py ./SSAKE -f CC57C_paired.fa -p 1 -g CC57C_unpaired.fa -m 20 -w 5 -b run2014
./ssake_v4.0/tools/getStats.pl run2014_contigs.fa
./ssake_v4.0/tools/getStats.pl run2014_scaffolds.fa

TRY IT OUT BY SIMPLY RUNNING:

cd test;./MiSeqCampylobacterAssemblyPIPELINE.sh
cd test;./MiSeqCampylobacterAssembly.sh

What's new in v3.7+ ?


v3.7+ Improved support for seed-based -s assemblies, notably read-space restriction option -u (TASR behavior, without fastq support)

What's new in v3.6+ ?


v3.6+ supports various insert size sequence libraries.  To work with paired data, users must add ":insert_size" (e.g. >SLXA23-1-1-2-13:200 for a 200bp library) at the end of the fasta header (>) for each pairs. v3.6 also has preliminary support for Sanger, paired-end reads.

What's new in v3.5+ ?


In v3.5, the read pairing logic is used in the extension process.  More specifically, passed the upper bound insert size, only forward reads AND reverse reads having their assembled mate already assembled in the contig being built will be considered for extension.  This has for effect to mitigate contig misassemblies due to repeats.  It will also extend the end of adjacent contigs in a scaffold in an effort to fill gaps - the resulting contigs are placed in the .mergedcontigs file.

What's new in v3.4+ ?


Version 3.4 exploits paired-end reads to explore possible contig merges within scaffolds (Consecutive contigs >= -z bases must overlap by -m bases or more). Version v3.4.1 allows a user to merge all contigs of a scaffold by padding predicted gaps with Ns (-n 1) and predicted but undetected overlaps with a single (n).  Merged contigs are outputted in the _mergedcontigs.fa file.  The default behaviour (-n 0) is to NOT pad the gaps with Ns (v3.4 behaviour).  In the v3.4.1, the _readposition.txt file tracks read names instead of read sequences as the latter can be inferred from the start and end coordinates. THIS IS NOW (v4.0) DEPRECATED.

What's new in v3.3+ ?


Fixed a bug in PET routine.  User can now track read position and individual base coverage for reads *fully embedded* within contigs, using the -c option.

What's new in v3.2.1+ ?


This release runs ~30% faster and requires ~33% less RAM compared to 15-node prefix tree SSAKE v3.2.0.1 beta.
Compared to the previous v3.2 release (11-node prefix tree), it will run at ca. double the speed, requiring ~20% more RAM.

What's new in v3.2+ ?


The -t option, first introduced in SSAKE 1.3, is back in this release.  The option allows you to trim your contigs in 3', 1 base at a time until a maximum base trim value (-t) is reached.  This option yields longer contigs, but increases assembly run time and, at high t values, might introduce contig misassemblies if your run parameters (i.e. -m, -o & -r) are not stringent enough.  At -m 16 -o 3 -r 0.7, best results were obtained with -t 1.  That's because it removes bases that cause premature breaks during the fragment assembly.  If set, end-trimming kicks in only when all possibilities have been exhausted for a contig extension.
This release also fixes a major bug that prevented SSAKE from exploring the entire read space for contig extensions seeded by the shorter reads.

What's new in v3.1+ ?


SSAKE now allows users to input a fasta file with DNA sequences for use as seeds to nucleate contig extension.
This feature can be used to extend existing/known DNA sequences using millions of short reads.
There's a new input format for paired-end reads, which allows reads of variable length to be considered (such as quality-trimmed reads)

What's new in v3.0+ ?


SSAKE supports Illumina paired-end read data to build scaffolds. 

What's new in v2.0+ ?


SSAKE can now handle error-rich data sets by looking through the overlapping k-mer space for consensus bases overhanging a seed sequence or contig.
SSAKE now runs on reads of various lengths.  That means quality base trimming of individual sequences can be achieved (using TQS.py/TQSfastq.py supplied in ./tools directory).

Implementation and requirements


SSAKE is implemented in PERL and runs on any OS where PERL is installed.

Side-by-side comparison between ssake2.0 and vcake1.0 indicates that SSAKE is nearly 3-fold faster and yields contigs that are as contiguous and accurate.

The python version 2.0 (released in ssake_v2.0.tar.gz and distributed under ./tools) has not yet been fully tested. Due to SSAKE's memory requirements, you would need a version of the perl interpreter compiled for 64-bit computers if you intend to assemble millions of short sequences. Development of SSAKE was done using perl v5.8.5 built for x86_64-linux-thread-multi

With ssake_3.2.pl -p 0 You can assemble ~5 million 25-mers with SSAKE on a computer with 4GB RAM
You can assemble 60-80 million 25-mers with SSAKE on a computer with 32GB RAM

PLEASE READ: *When using paired-end reads (-p 1), SSAKE tracks in memory all paired reads located in contigs >= z. That means that the memory usage will increase drastically with the size of your data set. Just be aware of this limitation and don't be surprised if you observe a lot of data swapping to disk if you attempt to run SSAKE on a machine with little RAM.

SSAKE might not be suited to work with 454-type reads. Simply because recurring base insertions/deletions errors, such as those commonly seen in homopolymeric regions, will not assemble well in the context of the SSAKE algorithm scheme. Sanger reads are ok, as long as reads are quality-trimmed.

Installation


Download the tar ball, gunzip and extract the files on your system using:

gunzip ssake_v4-0.tar.gz
tar -xvf ssake_v4-0.tar

Change the shebang line of SSAKE to point to the version of perl installed on your system and you're good to go.

To install this package with Conda, run one of the following:

conda install -c bioconda ssake
conda install -c "bioconda/label/cf201901" ssake

Documentation


Refer to the SSAKE.readme file on how to run SSAKE and the SSAKE web site for information about the software and its performance

Questions or comments? We would love to hear from you!

Citing SSAKE


Thank you for your Stars and for using, developing and promoting this free software!

If you use SSAKE (or any other tools distributed with SSAKE) for you research, please cite:

Assembling millions of short DNA sequences using SSAKE
Warren RL, Sutton GG, Jones SJM, Holt RA.  
2007. Bioinformatics. 23(4):500-501

link

Running SSAKE


e.g. SSAKE -f paired.fa -m 20 -o 3 -r 0.7 -p 1 -c 1 -e 0.75 -l 5 -a 0.3 -z 50 -w 5 -g unpaired.fa 

Usage: ./SSAKE [v4.0]
-f  File containing all the [paired (-p 1)] reads (required)
          With -p 1:
        ! Target insert size must be indicated at the end of the header line (e.g. :400 for a 400bp fragment/insert size)
        ! Paired reads must be separated by ":"
          >header:400 (or >header_barcode:400)
          ACGACACTATGCATAAGCAGACGAGCAGCGACGCAGCACG:GCGCACGACGCAGCACAGCAGCAGACGAC
-g  Fasta file containing unpaired sequence reads (optional)
-w  Minimum depth of coverage allowed for contigs (e.g. -w 1 = process all reads [v3.7 behavior], required, recommended -w 5)
    *The assembly will stop when 50+ contigs with coverage < -w have been seen.*
-s  Fasta file containing sequences to use as seeds exclusively (specify only if different from read set, optional)
        -i Independent (de novo) assembly  i.e Targets used to recruit reads for de novo assembly, not guide/seed reference-based assemblies (-i 1 = yes (default), 0 = no, optional)
        -j Target sequence word size to hash (default -j 15)
        -u Apply read space restriction to seeds while -s option in use (-u 1 = yes, default = no, optional)
-m  Minimum number of overlapping bases with the seed/contig during overhang consensus build up (default -m 20)
-o  Minimum number of reads needed to call a base during an extension (default -o 2)
-r  Minimum base ratio used to accept a overhang consensus base (default -r 0.7)
-t  Trim up to -t base(s) on the contig end when all possibilities have been exhausted for an extension (default -t 0, optional)
-c  Track base coverage and read position for each contig (default -c 0, optional)
-y  Ignore read mapping to consensus (-y 1 = yes, default = no, optional)
-h  Ignore read name/header *will use less RAM if set to -h 1* (-h 1 = yes, default = no, optional)
-b  Base name for your output files (optional)
-z  Minimum contig size to track base coverage and read position (default -z 100, optional)
-q  Break tie when no consensus base at position, pick random base (-q 1 = yes, default = no, optional)
-p  Paired-end reads used? (-p 1 = yes, default = no, optional)
-v  Runs in verbose mode (-v 1 = yes, default = no, optional)
============ scaffolding options below only considered with -p 1 ============
-e  Error (%) allowed on mean distance   e.g. -e 0.75  == distance +/- 75% (default -e 0.75, optional)
-l  Minimum number of links (read pairs) to compute scaffold (default -k 5, optional)
-a  Maximum link ratio between two best contig pairs *higher values lead to least accurate scaffolding* (default -a 0.3, optional)

A BASIC PIPELINE TO HELP YOU PREPARE YOUR INPUT PAIRED READS EXISTS (./test/runSSAKE.sh)

USAGE: ./runSSAKE.sh read1.fq read2.fq libraryFragmentLength basename


Test data


Go to the test folder, (cd test)

run ../SSAKE -f Herpesvirus_3.60kb.reads.fa -m 16 -o 2 -r 0.6 -p 0 -t 0 -c 1 -w 5 -b myFirstSSAKErun compare your results with Herpesvirus_3.60kb.reference.fa ** This error-rich, simulated data set was made available as part of the VCAKE v1.0 distribution it represents a 60,000 bp stretch of the NC_001348.1 human herpes virus 3 **

run ../SSAKE -f Herpesvirus_3.60kb.reads.fa -m 16 -o 2 -w 5 -b seedtest -c 1 -s Herpesvirus_3.60kb.seed.fa -u 1 -i 0 -j 20 The assembly above uses a single seed sequence located in (Herpesvirus_3.60kb.seed.fa) compare your results with Herpesvirus_3.60kb.reference.fa to see how successful the assembly was

TEST DATA / SSAKE ASSEMBLIES

A) Testing the distribution with very short reads:
../SSAKE -f Herpesvirus_3.60kb.reads.fa -m 16 -o 2 -w 5 -c 1 

B) Testing the targeted assembly using a seed/target sequence:
../SSAKE -f Herpesvirus_3.60kb.reads.fa -m 16 -o 2 -w 5 -b seedtest -c 1 -s Herpesvirus_3.60kb.seed.fa -u 1 -i 0 -j 20

C) Testing SSAKE on real (experimental) Illumina sequence data


1) Ebola (Zaire ebolavirus isolate Ebola)

./MiSeqEbolaAssemblyPIPELINE.sh
(read download,trimming,formatting,assembly)


2) Campylobacter showae - CRC tumor isolate / Illumina MiSeq data
./MiSeqCampylobacterAssemblyPIPELINE.sh
(read download,trimming,formatting,assembly)

or 

./MiSeqCampylobacterAssembly.sh
(read download, assembly)


3) Escherichia coli / 2014 Illumina MiSeq data

./MiSeqEcoliAssembly250XPE300.sh
(read download, assembly)

This is illumina MiSeq base space data (one tenth of 2500-fold coverage run)
sequence ~ 250X, 550bp fragments PE300

*compare your assembly to:
coliMiSeq300m80.contigs.stats1
coliMiSeq300m80.scaffolds.stats1


4) Fusobacterium nucleatum - CRC tumor isolate / Illumina HiSeq 2000 data
./HiSeqFusobacteriumAssembly.sh
(read download, assembly)


5) De Novo Targeted assembly of a TMPRSS2:ERG fusion using a prostate adenocarcinoma RNA-seq dataset
./runSSAKEtargeted.sh
(read download,trimming,formatting,assembly)


D) Testing SSAKE on LRsim linked-read data

1) C. elegans linked-read assembly
./CelegansLinkedReadsAssembly.sh
(read download, assembly)

SSAKE v4.0 RUN TESTS (all tests provided in the test folder) January 2018

Contigs:

species read type n n:500 L50 min N80 N50 N20 E-size max sum time (h:mm:ss) RAM (GB) file name
Herpes virus Illumina single-end 7 1 1 60,000 60,000 60,000 60,000 60,000 60,000 60,000 0:00:22 0.2 HStestInstall_contigs.fa
Ebola virus Illumina paired-end 122 18 1 501 688 18,917 18,917 12,139 18,917 30,101 0:01:22 0.4 ebola_contigs.fa
F. nucleatum HiSeq paired-end 697 560 131 501 2,540 4,641 8,019 5,631 19,764 1,937,271 0:10:43 6.0 fusoCC53_contigs.fa
C. showae MiSeq paired-end 206 109 21 562 18,061 32,736 64,004 40,613 107,028 2,222,124 0:06:58 3.5 CC57C_contigs.fa
E. coli MiSeq paired-end 101 97 12 687 80,150 127,399 210,819 146,845 318,067 4,606,049 0:21:50 9.3 coliMiSeq300m80_contigs.fa
C. elegans LRsim linked-read (paired) 5,352 4,949 613 502 19,304 44,626 90,857 58,011 305,563 95.57e6 4:05:34 79.1 celegansLR_contigs.fa

Scaffolds:

species read type n n:500 L50 min N80 N50 N20 E-size max sum name
F. nucleatum HiSeq paired-end 638 524 120 501 2,756 5,067 9,273 6,141 19,764 1,943,677 fusoCC53_scaffolds.fa
C. showae MiSeq paired-end 135 42 6 585 47,914 128,053 233,500 159,643 382,374 2,223,249 Cshowae_scaffolds.fa
E. coli MiSeq paired-end 86 82 5 687 91,172 204,889 1,193,524 437,907 1,193,524 4,606,049 coliMiSeq300m80_scaffolds.fa
C. elegans LRsim linked-read (paired) 4,637 4,309 491 502 22,274 53,866 112,496 74,063 357,928 95.59e6 celegansLR_scaffolds.fa

stats generated with abyss-fac

benchmark: Intel(R) Xeon(R) CPU E7-8867 v3 @ 2.50GHz 128CPU 2TB RAM CentOS7/ 1 thread per assembly

Algorithm


  1. Sequence Overlap

Short DNA sequences of length l in a single multi fasta file -f are read in memory, populating a hash table keyed by unique sequence reads with pairing values representing the number of sequence occurrence in the input read set. The normalized sequence reads are sorted by decreasing abundance (number of times the sequence is repeated) to reflect coverage and minimize extension of reads containing sequencing errors. Reads having sequencing errors are more likely to be unique in the entire read set when compared to their error-free counterparts. Sequence assembly is initiated by generating the longest 3'-most word (k-mer) from the unassembled read u that is shorter than the sequence read length l. Every possible 3' most k-mers will be generated from u and used in turn for the search until the word length is smaller than a user-defined minimum, m. Meanwhile, all perfectly overlapping reads will be collected in an array and further considered for 3' extension once the k-mer search is done. At the same time, a hash table c will store every base along with a coverage count for every position of the overhang (or stretches of bases hanging off the seed sequence u).

Once the search complete, a consensus sequence is derived from the hash table c, taking the most represented base at each position of the overhang. To be considered for the consensus, each base has to be covered by user-defined -o (set to 2 by default). If there's a tie (two bases at a specific position have the same coverage count), the prominent base is below a user-defined ratio r, the coverage -o is to low or the end of the overhang is reached, the consensus extension terminates and the consensus overhang joined to the seed sequence/contig. All reads overlapping are searched against the newly formed sequence and, if found, are removed from the hash table and prefix tree. If they are not part of the consensus, they will be used to seed/extend other contigs, if applicable. If no overlapping reads match the newly formed contig, the extension is terminated from that end and SSAKE resumes with a new seed. That prevents infinite looping through low-complexity DNA sequences. In the former case, the extension resumes using the new [l-m] space to search for joining k-mers.

The process of progressively cycling through longer to shorter 3'-most k-mer is repeated after every sequence extension until nothing else can be done on that side. Since only left-most searches are possible with a prefix tree, when all possibilities have been exhausted for the 3' extension, the complementary strand of the contiguous sequence generated is used to extend the contig on the 5' end. The DNA prefix tree is used to limit the search space by segregating sequence reads and their reverse-complemented counterparts by their first eleven 5' end bases.

There are three ways to control the stringency in SSAKE:

i) Disallow read/contig extension if the coverage is too low (-o).  Higher -o values lead to shorter contigs, but minimizes sequence misassemblies.
ii) Adjust the minimum overlap -m allowed between the seed/contig and short sequence reads.  Higher m values lead to more accurate contigs at the cost of decreased contiguity.  
iii) Set the minimum base ratio -r to higher values
  1. Building scaffolds with SSAKE

If the -p option is set to 1, it is assumed that the data supplied in the fasta file (-f) consists of paired-end reads, concatenated together on the same line, but separated by ":" -- see "Input sequences" section below.

During data input, pairs are split and both used to fill the prefix tree and hash table, as described in Warren et al. 2007. With the -p option set, the position of all sequence reads in contigs -z and larger are tracked. If a file is specified with -g, its unpaired sequences will be co-assembled along with paired reads during the SSAKE 3' extension but the former reads will NOT be tracked.

At the end of the overlap phase (aka contig extension), the -f fasta file is read again, associating reads with their pairs. For each read pairs, putative contig pairs (pre-scaffolding stage) are tallied based on the position/location of the paired-end reads on different contigs. Contig pairs are only considered if the calculated distance between them satisfy the mean distance specified (>template:insert_size) while allowing for a deviation (-e), also defined by the user. Only contig pairs having a valid gap or overlap are allowed to proceed to the scaffolding stage. Please note that this stage accepts redundancy of contig pairs (i.e. a given contig may link to multiple contigs, and the number of links (spanning pairs) between any given contig pair is recorded, along with a mean putative gap or overlap(-)). Once pairing between contigs is complete, the scaffolds are built using contigs (-z or larger) as seeds. Every contig is used in turn until all have been incorporated into a scaffold.

Consider the following contig pairs (AB, AC and rAD):

    A         B
========= ======== 
  ->       <-
   ->        <-
    ->      <-
       ->       <-

    A       C
========= ======
  ->        <-
    ->        <-

   rA        D           equivalent to rDA, in this order
========= =======
      ->   <-
     ->   <-
       ->   <-

Two parameters control scaffolding (-k and -a). The -k option specifies the minimum number of links (read pairs) a valid contig pair MUST have to be considered. The -a option specifies the maximum ratio between the best two contig pairs for a given seed/contig being extended. For example, contig A shares 4 links with B and 2 links with C, in this orientation. contig rA (reverse) also shares 3 links with D. When it's time to extend contig A (with the options -k and -a set to 2 and 0.7, respectively), both contig pairs AB and AC are considered. Since C (second-best) has 2 links and B (best) has 4 (2/4) = 0.5 below the maximum ratio of 0.7, A will be linked with B in the scaffold and C will be kept for another extension. If AC had 3 links the resulting ratio (0.75), above the user-defined maximum 0.7 would have caused the extension to terminate at A, with both B and C considered for a different scaffold. A maximum links ratio of 1 (not recommended) means that the best two candidate contig pairs have the same number of links -- SSAKE will accept the first one since both have a valid gap/overlap. When a scaffold extension is terminated on one side, the scaffold is extended on the "left", by looking for contig pairs that involve the reverse of the seed (in this example, rAD). With AB and AC having 4 and 2 links, respectively and rAD being the only pair on the left, the final scaffolds outputted by SSAKE would be:

  1. rD-A-B
  2. C

SSAKE outputs a .scaffolds file with linkage information between contigs (see "Understanding the .scaffolds csv file" below) Accurate scaffolding depends on many factors. Number and nature of repeats in your target sequence, optimum adjustments of insert_size, -e, -k and -a and data quality/size of sequence set (more doesn't mean better) will all affect SSAKE's ability to build scaffolds.

  1. Using a seed sequence file

If the -s option is set and points to a valid fasta file, the DNA sequences comprised in that file will populate the hash table and be used exclusively as seeds to nucleate contig extensions (they will not be utilized to build the prefix tree). In that scheme, every unique seed will be used in turn to nucleate an extension, using short reads found in the tree (specified in -f). This feature might be useful if you already have characterized sequences & want to increase their length using short reads. That said, since the short reads are not used as seeds when -s is set, they will not co-assemble with one another WITHOUT a seed sequence file - unless you run SSAKE in targeted de novo assembly mode (see below). Also, to speed up the assembly, no imbedded reads (i.e. those aligning to the seed in their entirety) are considered. Only reads that contribute to extending a seed sequence are noted.

When -s is set, the _contigs.fa file lists all extended seeds, even if it's by a single base. The .singlets will ONLY list seeds that could not be extended. Unassembled microreads will NOT be outputted.

Support for sequence target-independent de novo assemblies:

The -i option instructs SSAKE/TASR to use target sequences for the sole purpose of recruiting sequence reads. If set (-i 1) the target sequences will not seed de novo assemblies and this task will be achieved by recruited reads in a target-independent fashion instead. This has the advantage of allowing the user to provide, as a target, a large reference sequence (-s) without a priori knowledge of variant bases or other structural variants.

SSAKE v3.8.2+/TASR v1.5+ no longer constrains the use of 15-character words derived from a target sequence for interrogating candidate reads. User-defined target word length values are now passed to the algorithm using the -j option. Using larger -j values should help speed up the search when using long sequence reads, since it will restrict the sequence space accordingly. Note: whereas specificity, speed and RAM usage may increase with -j, it may yield more sparse/fragmented assemblies. Proper experimentation with various -j values are warranted.

*Refer to the "Test data" section below for a concrete example

  1. Using seeds (-s) with mate pairs (-p 1):

If more than one seed is supplied in the -s file and you're providing paired-end reads (-p 1), SSAKE will attempt to scaffold extended seeds (if a seed wasn't extended it will end-up in the singlets and will not be considered for scaffolding) using the supplied mate pairs.

Input


UNPAIRED:

DNA sequences can be in lower caps as well (NO UNDERSCORE _ CHARACTER ALLOWED)

>PX1CG29
TTAACACTTTCGGATATTTCTGATG
>PX1CG35
CTTTCGGATATTTCTGATGAGTCGA
>PX1CG64
TTATCTTGATAAAGCAGGAATTACT
...

PAIRED:

>2-1-464-197:200
TGGCTCACCCCTGTAATCCCAGCACT:CTCCCAGGTTCAAGCGATTCTCCTGC
>2-1-783-425:300
GTCTGAGGGTCCCAGGAACCAG:TGCCCCAGAGGTGGGAGCAGGGGA
>2-1-662-655:1000
TGAATCCCCACCAGGCGCCTTCGG:CACTTTATTATTAATGTACAAAAT
...

PAIRED, LINKED READ:

>2-1-464-197_ACGATGCATGCAGTAG:200
TGGCTCACCCCTGTAATCCCAGCACT:CTCCCAGGTTCAAGCGATTCTCCTGC
>2-1-783-425_ACGATGCATGCAGTAG:300
GTCTGAGGGTCCCAGGAACCAG:TGCCCCAGAGGTGGGAGCAGGGGA
>2-1-662-655_ATGCATGCATGCTAGC:1000
TGAATCCCCACCAGGCGCCTTCGG:CACTTTATTATTAATGTACAAAAT
...

-Note: For linked read, a barcode sequence of any length will do. In fact, any [A-Za-z0-9] characters after the underscore, before : or /\ will be used as barcode sequence to filter the reads. Note that when linked reads are not used, but regular reads are used instead, the original ssake behaviour will take precedence. Note that under NO circumstances should your read contain underscore characters (_) if your reads are NOT linked.

Example input from the C. elegans test data available with the v4.0 distribution:

>gi|453232919|ref|NCI003284.9|I11747444I11747218I1I0I0I0I0:0:0I0:0:0I98920/1_AAACCTGAGCTTTCAG/:350
CACATACGAGGGCGTTATTTGAAAAATTTAAAAATCAACATGTTCAAGCGTGCGAAGTGTCAAAATAAAAAAGAAAAAAAAAACGAAAAAAAAAACAGAAAAGGCTGATAAGAGGACGCGTCAAGTTA:ACTGCTCATTTGTCAATCAGCAAGGTACATGAAAACACAGAGCAGGAACCAAAATGCACACAATAAAACTCCCCGTACCCATTGTGTGGTACGCAGTACAAAATGACTGACAATAAGAAAGGGAGAGAGGGATTGAGGCGCCGAATACTTG

Additional info:

-Paired sequences must be concatenated together in one fasta-like entry, separated by ":".  For example, TGGCTCACCCCTGTAATCCCAGCACT:CTCCCAGGTTCAAGCGATTCTCCTGC consists of two paired reads.  Changes to the input was made to allow reads of variable length (e.g. quality-trimmed reads) to be considered by SSAKE.  As of v3-6, the header line [>] must have [:insert_size] at the very end (see above example)

-The -f option can read either paired or unpaired sequences, depending whether -p is set or not, respectively.
Users can co-assemble paired and unpaired reads if they wish.  If so, the unpaired reads are inputted using the -g option.

General points:

-To be considered, sequences have to be longer than 16 nt or -m (but can be of different lengths).  If they are shorter, the program will simply omit them from the assembly and will be placed in the .shorts file 
-Short sequences that have not been extended are placed in the .singlets file
-As before, the length of individual sequence is used to determine the size of the right-most subsequence to look for initially
-Reads containing ambiguous bases "." and characters other than ACGT will be ignored entirely
-Spaces in fasta file are NOT permitted and will either not be considered or result in execution failure

Output


Output file (-p 0 and -p 1) Description
_contigs.fa fasta file; All sequence contigs
_scaffolds.fa fasta file; All sequence scaffolds
.log text file; Logs execution time / errors / pairing stats (if -p is set to 1)
_short.txt text file; Lists sequence reads shorter than a set, acceptable, minimum
_singlets.fa fasta file; All unassembled sequence reads
Output file (-p 1) Description
_pairing-distribution.csv comma-separated file; 1st column is the calculated distance for each pair (template) with reads that assembled logically within the same contig. 2nd column is the number of pairs at that distance
_pairing-issues.txt text file; Lists all pairing issues encountered between contig pairs and illogical/out-of-bounds pairing
.scaffolds comma-separated file; see below
Output file (-c 1*) Description
_readposition.txt this is a text file listing all whole (fully embedded) reads, start and end coordinate onto the contig (in this order). For reads aligning on the minus strand, end coordinate is < start coordinate
_coverage.csv this is a comma-separated values file showing the base coverage at every position for any given contig > -z
*WARNING: ASSOCIATED FILES CAN BECOME VERY LARGE!

Understanding the _contigs.fa fasta header


e.g.
>contig27|size52|read193|cov92.79

contig id# = 27 size (G) = 52 nt number of reads (N) = 193 cov [coverage] (C) = 92.79

the coverage (C) is calculated using the total number (T) of consensus bases [sum(L)] provided by the assembled sequences divided by the contig size:

C = T / G

Understanding the .scaffolds layout csv file


e.g.
scaffold1,7484,f127Z7068k12a0.58m42_f3090z62k7a0.14m76_f1473z354

column 1: a unique scaffold identifier column 2: the sum of all contig sizes that made it to the scaffold/supercontig column 3: a contig chain representing the layout:

e.g. f127Z7068k12a0.58m42_f3090z62k7a0.14m76_f1473z354

means: contig f127 (strand=f/+), size (z) 7068 (Z if contig was used as the seed sequence) has 12 links (k), link ratio of 0.58 (a) with a mean gap of 42nt (m) with reverse (r) of contig 3090 (size 62) on the right. if m values are negative, it's just that a possible overlap was calculated using the mean distance supplied by the user and the position of the reads flanking the contig. Negative m values imply that there's a possible overlap between the contigs. But since the pairing distance distribution usually follows a Normal/Gaussian distribution, some distances are expected to be larger than the median size expected/observed. In reality, if the exact size was known between each paired-reads, we wouldn't expect much negative m values unless a break occurred during the contig extension (likely due to base errors/SNPs).

Use makeFastaFileFromScaffolds.pl included in this distribution to make a scaffold fasta file (ordered and oriented contig sequences) using the layout recipe (contig chain) shown above.

Understanding the _coverage.csv file


e.g.
>contig1|size60000|read74001|cov37.00
12,12,13,13,13,14,14,15,16,16,20,21,22,23,25,26,27,28,27 ...

Each number represents the number of reads covering that base at that position.

Understanding the _readposition.txt file


e.g.
>contig1|size60000|read74001|cov37.00
READ_85952,3,32
READ_92647,6,35
READ_72602,8,37
READ_29659,9,38
READ_74582,11,40
READ_97793,11,40
READ_85742,11,40
READ_95375,12,41
READ_9721,15,44
READ_49141,16,45
READ_43328,18,1
READ_94449,18,1

In this order: read name [template th -p 1 :: name followed with 1 or 2, corresponds to the order in the sequence input (1:2)], start coordinate, end coordinate. end < start indicates read is on minus strand

SSAKE does not


  1. Take into consideration base quality scores. It is up to the user to process the sequence data before assembling with SSAKE.
  2. Consider sequence read having any character other than A,C,G,T and will skip these reads entirely while reading the fasta file.

Note: Python scripts (TQS.py, TQSfastq.py, TQSexport.fq) are provided to help trim poor quality bases off Illumina sequences. Refer to TQS.readme and TRIMMING_PAIRED_READS.README included in this distribution (in the ./tools subdirectory) for information on how to run those programs

License


SSAKE Copyright (c) 2006-present Canada's Michael Smith Genome Science Centre. All rights reserved.

This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

Enjoy SSAKE responsibly!