1. Co-assembly procedure with read mapping for estimation of the abundances of genes in each metagenome
2. Co-assembly of a large number of metagenomes via merging of individual metagenomes
3. Includes binning and bin checking, for retrieving individual genomes
4. The results are stored in a database, where they can be easily exported and shared, and can be inspected anywhere using a web interface.
5. Internal checks for the assembly and binning steps inform about the consistency of contigs and bins, allowing to spot potential chimeras.
6. Metatranscriptomic support via mapping of cDNA reads against reference metagenomes

Address of the bookmark: https://github.com/jtamames/SqueezeMeta

Manual at http://spades.bioinf.spbau.ru/release3.7.1/manual.html

Address of the bookmark: http://bioinf.spbau.ru/spades

Cleaning your data in this way is often required: Reads from small-RNA sequencing contain the 3’ sequencing adapter because the read is longer than the molecule that is sequenced. Amplicon reads start with a primer sequence. Poly-A tails are useful for pulling out RNA from your sample, but often you don’t want them to be in your reads.

Cutadapt helps with these trimming tasks by finding the adapter or primer sequences in an error-tolerant way. It can also modify and filter reads in various ways. Adapter sequences can contain IUPAC wildcard characters. Also, paired-end reads and even colorspace data is supported. If you want, you can also just demultiplex your input data, without removing adapter sequences at all.

Cutadapt comes with an extensive suite of automated tests and is available under the terms of the MIT license.

If you use cutadapt, please cite DOI:10.14806/ej.17.1.200 .

Address of the bookmark: https://cutadapt.readthedocs.io/en/stable/installation.html#quickstart

Wgsim is a small tool for simulating sequence reads from a reference genome. It is able to simulate diploid genomes with SNPs and insertion/deletion (INDEL) polymorphisms, and simulate reads with uniform substitution sequencing errors. It does not generate INDEL sequencing errors, but this can be partly compensated by simulating INDEL polymorphisms.

Wgsim outputs the simulated polymorphisms, and writes the true read coordinates as well as the number of polymorphisms and sequencing errors in read names. One can evaluate the accuracy of a mapper or a SNP caller with wgsim_eval.pl that comes with the package.

Address of the bookmark: https://github.com/lh3/wgsim

Address of the bookmark: http://bioinformatics.oxfordjournals.org/content/early/2014/02/12/bioinformatics.btu078.full.pdf

Genome sequences of rare, uncultured bacteria obtained by differential coverage binning of multiple metagenomes

Mads Albertsen, Philip Hugenholtz, Adam Skarshewski, Gene W. Tyson, Kåre L. Nielsen and Per .H. Nielsen

Nature Biotechnology 2013, doi: 10.1038/nbt.2579

Address of the bookmark: https://github.com/MadsAlbertsen/multi-metagenome

• Mapping Nanopore reads

BBMap.sh has a length cap of 6kbp. Reads longer than this will be broken into 6kbp pieces and mapped independently.

$ mapPacBio.sh -Xmx20g k=7 in=reads.fastq ref=reference.fa maxlen=1000 minlen=200 idtag ow int=f qin=33 out=mapped1.sam minratio=0.15 ignorequality slow ordered maxindel1=40 maxindel2=400

The "maxlen" flag shreds them to a max length of 1000; you can set that up to 6000. But I found 1000 gave a higher mapping rate.  

• Using Paired-end and single-end reads at the same time

BBMap itself can only run single-ended or paired-ended in a single run, but it has a wrapper that can accomplish it, like this:

$ bbwrap.sh in1=read1.fq,singletons.fq in2=read2.fq,null out=mapped.sam append

This will write all the reads to the same output file but only print the headers once. I have not tried that for bam output, only sam output

Note about alignment stats: For paired reads, you can find the total percent mapped by adding the read 1 percent (where it says "mapped: N%") and read 2 percent, then dividing by 2. The different columns tell you the count/percent of each event. Considering the cigar strings from alignment, "Match Rate" is the number of symbols indicating a reference match (=) and error rate is the number indicating substitution, insertion, or deletion (X, I, D).

• Exact matches when mapping small reads (e.g. miRNA)

When mapping small RNA's with BBMap use the following flags to report only perfect matches.

ambig=all vslow perfectmode maxsites=1000

It should be very fast in that mode (despite the vslow flag). Vslow mainly removes masking of low-complexity repetitive kmers, which is not usually a problem but can be with extremely short sequences like microRNAs.

• Important note about BBMap alignments

BBMap is always nondeterministic when run in paired-end mode with multiple threads, because the insert-size average is calculated on a per-thread basis, which affects mapping; and which reads are assigned to which thread is nondeterministic. The only way to avoid that would be to restrict it to a single thread (threads=1), or map the reads as single-ended and then fix pairing afterward:

bbmap.sh in=reads.fq outu=unmapped.fq int=f
repair.sh in=unmapped.fq out=paired.fq fint outs=singletons.fq

In this case you'd want to only keep the paired output. 

BBSplit is based on BBMap, so it is also nondeterministic in paired mode with multiple threads. BBDuk and Seal (which can be used similarly to BBSplit) are always deterministic. 

--------------------------------------------------------

Reformat.sh

• Count k-mers/find unknown primers

$ reformat.sh in=reads.fq out=trimmed.fq ftr=19

This will trim all but the first 20 bases (all bases after position 19, zero-based).

$ kmercountexact.sh in=trimmed.fq out=counts.txt fastadump=f mincount=10 k=20 rcomp=f

This will generate a file containing the counts of all 20-mers that occurred at least 10 times, in a 2-column format that is easy to sort in Excel. 

ACCGTTACCGTTACCGTTAC	100
AAATTTTTTTCCCCCCCCCC	85

...etc. If the primers are 20bp long, they should be pretty obvious.  

• Convert SAM format from 1.4 to 1.3 (required for many programs)

$ reformat.sh in=reads.sam out=out.sam sam=1.3

• Removing N basecalls

You can use BBDuk or Reformat with "qtrim=rl trimq=1". That will only trim trailing and leading bases with Q-score below 1, which means Q0, which means N (in either fasta or fastq format). The BBMap package automatically changes q-scores of Ns that are above 0 to 0 and called bases with q-scores below 2 to 2, since occasionally some Illumina software versions produces odd things like a handful of Q0 called bases or Ns with Q>0, neither of which make any sense in the Phred scale.

• Sampling reads

$ reformat.sh in=reads.fq out=sampled.fq sample=3000

To sample 10% of the reads:
reformat.sh in1=reads1.fq in2=reads2.fq out1=sampled1.fq out2=sampled2.fq samplerate=0.1

or more concisely:
reformat.sh in=reads#.fq out=sampled#.fq samplerate=0.1

and for exact sampling:
reformat.sh in=reads#.fq out=sampled#.fq samplereadstarget=100k

• Changing fasta headers

Remove anything after the first space in fasta header. 

 reformat.sh in=sequences.fasta out=renamed.fasta trd

"trd" stands for "trim read description" and will truncate everything after the first whitespace.

• Extract reads from a sam file

$ reformat.sh in=reads.sam out=reads.fastq

• Verify pairing and optionally de-interleave the reads

$ reformat.sh in=reads.fastq verifypairing

• Verify pairing if the reads are in separate files

$ reformat.sh in1=r1.fq in2=r2.fq vpair

If that completes successfully and says the reads were correctly paired, then you can simply de-interleave reads into two files like this:

$ reformat.sh in=reads.fastq out1=r1.fastq out2=r2.fastq

• Base quality histograms

$ reformat.sh in=reads.fq qchist=qchist.txt

That stands for "quality count histogram". 

• Filter SAM/BAM file by read length

$ reformat.sh in=x.sam out=y.sam minlength=50 maxlength=200

• Filter SAM/BAM file to detect/filter spliced reads

$ reformat.sh in=mapped.bam out=filtered.bam maxdellen=50

You can set "maxdellen" to whatever length deletion event you consider the minimum to signify splicing, which depends on the organism.
-------------------------------------------------------------
Repair.sh

• "Re-pair" out-of-order reads from paired-end data files

$ repair.sh in1=r1.fq.gz in2=r2.fq.gz out1=fixed1.fq.gz out2=fixed2.fq.gz outsingle=singletons.fq.gz

--------------------------------------------------------------
BBMerge.sh

BBMerge now has a new flag - "outa" or "outadapter". This allows you to automatically detect the adapter sequence of reads with short insert sizes, in case you don't know what adapters were used. It works like this:

$ bbmerge.sh in=reads.fq outa=adapters.fa reads=1m

Of course, it will only work for paired reads! The output fasta file will look like this:

>Read1_adapter
GATCGGAAGAGCACACGTCTGAACTCCAGTCACATCACGATCTCGTATGCCGTCTTCTGCTTG
>Read2_adapter
GATCGGAAGAGCACACGTCTGAACTCCAGTCACCGATGTATCTCGTATGCCGTCTTCTGCTTG

If you have multiplexed things with different barcodes in the adapters, the part with the barcode will show up as Ns, like this:

GATCGGAAGAGCACACGTCTGAACTCCAGTCACNNNNNNATCTCGTATGCCGTCTTCTGCTTG  

Note: For BBMerge with micro-RNA, you need to add the flag mininsert=17. The default is 35, which is too long for micro-RNA libraries. 

• Identifying adapters

If you have paired reads, and enough of the reads have inserts shorter than read length, you can identify adapter sequences with BBMerge, like this (they will be printed to adapters.fa):

$ bbmerge.sh in1=r1.fq in2=r2.fq outa=adapters.fa

-----------------------------------------------------------------

BBDuk.sh

Note: BBDuk is strictly deterministic on a per-read basis, however it does by default reorder the reads when run multithreaded. You can add the flag "ordered" to keep output reads in the same order as input reads

• Finding reads with a specific sequence at the beginning of read

$ bbduk.sh -Xmx1g in=reads.fq outm=matched.fq outu=unmatched.fq restrictleft=25 k=25 literal=AAAAACCCCCTTTTTGGGGGAAAAA

In this case, all reads starting with "AAAAACCCCCTTTTTGGGGGAAAAA" will end up in "matched.fq" and all other reads will end up in "unmatched.fq". Specifically, the command means "look for 25-mers in the leftmost 25 bp of the read", which will require an exact prefix match, though you can relax that if you want.

So you could bin all the reads with your known sequence, then look at the remaining reads to see what they have in common. You can do the same thing with the tail of the read using "restrictright" instead, though you can't use both restrictions at the same time.  

$ bbduk.sh in=reads.fq outm=matched.fq literal=NNNNNNCCCCGGGGGTTTTTAAAAA k=25 copyundefined

With the "copyundefined" flag, a copy of each reference sequence will be made representing every valid combination of defined letter. So instead of increasing memory or time use by 6^75, it only increases them by 4^6 or 4096 which is completely reasonable, but it only allows substitutions at predefined locations. You can use the "copyundefined", "hdist", and "qhdist" flags together for a lot of flexibility - for example, hdist=2 qhdist=1 and 3 Ns in the reference would allow a hamming distance of 6 with much lower resource requirements than hdist=6. Just be sure to give BBDuk as much memory as possible.

• Removing illumina adapters (if exact adapters not known)

If you're not sure which adapters are used, you can add "ref=truseq.fa.gz,truseq_rna.fa.gz,nextera.fa.gz" and get them all (this will increase the amount of overtrimming, though it should still be negligible). 

• Removing illumina control sequences/phiX reads

bbduk.sh in=trimmed.fq.gz out=filtered.fq.gz k=31 ref=artifacts,phix ordered cardinality

• Identify certain reads that contain a specific sequence

$ bbduk.sh in=reads.fq out=unmatched.fq outm=matched.fq literal=ACGTACGTACGTACGTAC k=18 mm=f hdist=2

Make sure "k" is set to the exact length of the sequence. "hdist" controls the number of substitutions allowed. "outm" gets the reads that match. By default this also looks for the reverse-complement; you can disable that with "rcomp=f".  

• Extract sequences that share kmers with your sequences with BBDuk

$ bbduk.sh in=a.fa ref=b.fa out=c.fa mkf=1 mm=f k=31

This will print to C all the sequences in A that share 100% of their 31-mers with sequences in B. 

• Extract sequences that contain N's with BBDuk

bbduk.sh in=reads.fq out=readsWithoutNs.fq outm=readsWithNs.fq maxns=0

If you have, say, 100bp reads and only want to separate reads containing all 100 Ns, change that to "maxns=99".

General notes for BBDuk.sh 

BBDuk can operate in one of 4 kmer-matching modes:
Right-trimming (ktrim=r), left-trimming (ktrim=l), masking (ktrim=n), and filtering (default). But it can only do one at a time because all kmers are stored in a single table. It can still do non-kmer-based operations such as quality trimming at the same time.

BBDuk2 can do all 4 kmer operations at once and is designed for integration into automated pipelines where you do contaminant removal and adapter-trimming in a single pass to minimize filesystem I/O. Personally, I never use BBDuk2 from the command line. Both have identical capabilities and functionality otherwise, but the syntax is different.

------------------------------------------------------------------

Randomreads.sh

• Generate random reads in various formats

$ randomreads.sh ref=genome.fasta out=reads.fq len=100 reads=10000

You can specify paired reads, an insert size distribution, read lengths (or length ranges), and so forth. But because I developed it to benchmark mapping algorithms, it is specifically designed to give excellent control over mutations. You can specify the number of snps, insertions, deletions, and Ns per read, either exactly or probabilistically; the lengths of these events is individually customizable, the quality values can alternately be set to allow errors to be generated on the basis of quality; there's a PacBio error model; and all of the reads are annotated with their genomic origin, so you will know the correct answer when mapping.

Bear in mind that 50% of the reads are going to be generated from the plus strand and 50% from the minus strand. So, either a read will match the reference perfectly, OR its reverse-complement will match perfectly.

You can generate the same set of reads with and without SNPs by fixing the seed to a positive number, like this:

$ randomreads.sh maxsnps=0 adderrors=false out=perfect.fastq reads=1000 minlength=18 maxlength=55 seed=5

$ randomreads.sh maxsnps=2 snprate=1 adderrors=false out=2snps.fastq reads=1000 minlength=18 maxlength=55 seed=5

[As of BBmap v. 36.59] rendomreads.sh gains the ability to simulate metagenomes. 

coverage=X will automatically set "reads" to a level that will give X average coverage (decimal point is allowed).

metagenome will assign each scaffold a random exponential variable, which decides the probability that a read be generated from that scaffold. So, if you concatenate together 20 bacterial genomes, you can run randomreads and get a metagenomic-like distribution. It could also be used for RNA-seq when using a transcriptome reference.

The coverage is decided on a per-reference-sequence level, so if a bacterial assembly has more than one contig, you may want to glue them together first with fuse.sh before concatenating them with the other references. 

• Simulate a jump library

You can simulate a 4000bp jump library from your existing data like this.

$ cat assembly1.fa assembly2.fa > combined.fa
$ bbmap.sh ref=combined.fa
$ randomreads.sh reads=1000000 length=100 paired interleaved mininsert=3500 maxinsert=4500 bell perfect=1 q=35 out=jump.fq.gz

--------------------------------------------------------------
Shred.sh

$ shred.sh in=ref.fasta out=reads.fastq length=200

The difference is that RandomReads will make reads in a random order from random locations, ensuring flat coverage on average, but it won't ensure 100% coverage unless you generate many fold depth. Shred, on the other hand, gives you exactly 1x depth and exactly 100% coverage (and is not capable of modelling errors). So, the use-cases are different. 
---------------------------------------------------------------
Demuxbyname.sh

• Demultiplex fastq files when the tag is present in the fastq read header (illumina)

$ demuxbyname.sh in=r#.fq out=out_%_#.fq prefixmode=f names=GGACTCCT+GCGATCTA,TAAGGCGA+TCTACTCT,...
outu=filename

"Names" can also be a text file with one barcode per line (in exactly the format found in the read header). You do have to include all of the expected barcodes, though.

In the output filename, the "%" symbol gets replaced by the barcode; in both the input and output names, the "#" symbol gets replaced by 1 or 2 for read 1 or read 2. It's optional, though; you can leave it out for interleaved input/output, or specify in1=/in2=/out1=/out2= if you want custom naming.

----------------------------------------------------------------

Readlength.sh

• Plotting the length distribution of reads

$ readlength.sh in=file out=histogram.txt bin=10 max=80000

That will plot the result in bins of size 10, with everything above 80k placed in the same bin. The defaults are set for relatively short sequences so if they are many megabases long you may need to add the flag "-Xmx8g" and increase "max=" to something much higher.

Alternatively, if these are assemblies and you're interested in continuity information (L50, N50, etc), you can run stats on each or statswrapper on all of them:

stats.sh in=file

or

statswrapper.sh in=file,file,file,file…

----------------------------------------------------------------
Filterbyname.sh

By default, "filterbyname" discards reads with names in your name list, and keeps the rest. To include them and discard the others, do this:

$ filterbyname.sh in=003.fastq out=filter003.fq names=names003.txt include=t

----------------------------------------------------------------
getreads.sh

If you only know the number(s) of the fasta/fastq record(s) in a file (records start at 0) then you can use the following command to extract those reads in a new file.

$ getreads.sh in= id=<number,number,number...> out=

The first read (or pair) has ID 0, the second read (or pair) has ID 1, etc.

Parameters:
in= Specify the input file, or stdin.
out= Specify the output file, or stdout.
id= Comma delimited list of numbers or ranges, in any order.
For example: id=5,93,17-31,8,0,12-13 
----------------------------------------------------------------
Splitsam.sh

• Splits a sam file into forward and reverse reads

splitsam.sh mapped.sam plus.sam minus.sam unmapped.sam
reformat.sh in=plus.sam out=plus.fq
reformat.sh in=minus.sam out=minus.fq rcomp

----------------------------------------------------------------
BBSplit.sh

BBSplit now has the ability to output paired reads in dual files using the # symbol. For example:

$ bbsplit.sh ref=x.fa,y.fa in1=read1.fq in2=read2.fq basename=o%_#.fq

will produce ox_1.fq, ox_2.fq, oy_1.fq, and oy_2.fq

You can use the # symbol for input also, like "in=read#.fq", and it will get expanded into 1 and 2.  

Added feature: One can specify a directory for the "ref=" argument. If anything in the list is a directory, it will use all fasta files in that directory. They need a fasta extension, like .fa or .fasta, but can be compressed with an additional .gz after that. Reason this is useful is to use BBSplit is to have it split input into one output file per reference file.


NOTE: 1 By default BBSplit uses fairly strict mapping parameters; you can get the same sensitivity as BBMap by adding the flags "minid=0.76 maxindel=16k minhits=1". With those parameters it is extremely sensitive.

NOTE: 2 BBSplit has different ambiguity settings for dealing with reads that map to multiple genomes. In any case, if the alignment score is higher to one genome than another, it will be associated with that genome only (this considers the combined scores of read pairs - pairs are always kept together). But when a read or pair has two identically-scoring mapping locations, on different genomes, the behavior is controlled by the "ambig2" flag - "ambig2=toss" will discard the read, "all" will send it to all output files, and "split" will send it to a separate file for ambiguously-mapped reads (one per genome to which it maps).

NOTE: 3 Zero-count lines are suppressed by default, but they should be printed if you include the flag "nzo=f" (nonzeroonly=false). 

NOTE: 4 BBSplit needs multiple reference files as input; one per organism, or one for target and another for everything else. It only outputs one file per reference file.

Seal.sh, on the other hand, which is similar, can use a single concatenated file, as it (by default) will output one file per reference sequence within a concatenated set of references. 
--------------------------------------------------------------
Pileup.sh

• To generate transcript coverage stats

$ pileup.sh in=mapped.sam normcov=normcoverage.txt normb=20 stats=stats.txt

That will generate coverage per transcript, with 20 lines per transcript, each line showing the coverage for that fraction of the transcript. "stats" will contain other information like the fraction of bases in each transcript that was covered. 

• To calculate physical coverage stats (region covered by paired-end reads) 

BBMap has a "physcov" flag that allows it to report physical rather than sequenced coverage. It can be used directly in BBMap, or with pileup, if you already have a sam file. For example:

$ pileup.sh in=mapped.sam covstats=coverage.txt

• Calculating coverage of the genome

Program will take sam or bam, sorted or unsorted.

$ pileup.sh in=mapped.sam out=stats.txt hist=histogram.txt

stats.txt will contain the average depth and percent covered of each reference sequence; the histogram will contain the exact number of bases with a each coverage level. You can also get per-base coverage or binned coverage if you want to plot the coverage. It also generates median and standard deviation, and so forth.

It's also possible to generate coverage directly from BBMap, without an intermediate sam file, like this:

$ bbmap.sh in=reads.fq ref=reference.fasta nodisk covstats=stats.txt covhist=histogram.txt

We use this a lot in situations where all you care about is coverage distributions, which is somewhat common in metagenome assemblies. It also supports most of the flags that pileup.sh supports, though the syntax is slightly different to prevent collisions. In each case you can see all the possible flags by running the shellscript with no arguments.

• To bin aligned reads

$ pileup.sh in=mapped.sam out=stats.txt bincov=coverage.txt binsize=1000

That will give coverage within each bin. For read density regardless of read length, add the "startcov=t" flag.  

--------------------------------------------------------------
Dedupe.sh

Dedupe ensures that there is at most one copy of any input sequence, optionally allowing contaminants (substrings) to be removed, and a variable hamming or edit distance to be specified. Usage:

$ dedupe.sh in=assembly1.fa,assembly2.fa out=merged.fa

That will absorb exact duplicates and containments. You can use "hdist" and "edist" flags to allow mismatches, or get a complete list of flags by running the shellscript with no arguments.  

Dedupe will merge assemblies, but it will not produce consensus sequences or join overlapping reads; it only removes sequences that are fully contained within other sequences (allowing the specified number of mismatches or edits).

Dedupe can remove duplicate reads from multiple files simultaneously, if they are comma-delimited (e.g. in=file1.fastq,file2.fastq,file3.fastq). And if you set the flag "uniqueonly=t" then ALL copies of duplicate reads will be removed, as opposed to the default behavior of leaving one copy of duplicate reads.

However, it does not care which file a read came from; in other words, it can't remove only reads that are duplicates across multiple files but leave the ones that are duplicates within a file. That can still be accomplished, though, like this:

1) Run dedupe on each sample individually, so now there are at most 1 copy of a read per sample.
2) Run dedupe again on all of the samples together, with "uniqueonly=t". The only remaining duplicate reads will be the ones duplicated between samples, so that's all that will be removed.  

--------------------------------------------------------------

• Generate ROC curves from any aligner

[*]index the reference

$ bbmap.sh ref=reference.fasta

[*]Generate random reads

$ randomreads.sh reads=100000 length=100 out=synth.fastq maxq=35 midq=25 minq=15

[*]Map to produce a sam file

...substitute this command with the appropriate one from your aligner of choice

$ bbmap.sh in=synth.fq out=mapped.sam

[*]Generate ROC curve

$ samtoroc.sh in=mapped.sam reads=100000

--------------------------------------------------------------

• Calculate heterozygous rate for sequence data

$ kmercountexact.sh in=reads.fq khist=histogram.txt peaks=peaks.txt

You can examine the histogram manually, or use the "peaks" file which tells you the number of unique kmers in each peak on the histogram. For a diploid, the first peak will be the het peak, the second will be the homozygous peak, and the rest will be repeat peaks. The peak caller is not perfect, though, so particularly with noisy data I would only rely on it for the first two peaks, and try to quantify the higher-order peaks manually if you need to (which you generally don't).

-----------------------------------------------------------------

• Compare mapped reads between two files

To see how many mapped reads (can be mapped concordant or discordant, doesn't matter) are shared between the two alignment files and how many mapped reads are unique to one file or the other.

$ reformat.sh in=file1.sam out=mapped1.sam mappedonly
$ reformat.sh in=file2.sam out=mapped2.sam mappedonly

That gets you the mapped reads only. Then:

$ filterbyname.sh in=mapped1.sam names=mapped2.sam out=shared.sam include=t

...which gets you the set intersection;

$ filterbyname.sh in=mapped1.sam names=mapped2.sam out=only1.sam include=f
$ filterbyname.sh in=mapped2.sam names=mapped1.sam out=only2.sam include=f

...which get you the set subtractions.  

--------------------------------------------------------------

BBrename.sh

$ bbrename.sh in=old.fasta out=new.fasta

That will rename the reads as 1, 2, 3, 4, ... 222.

You can also give a custom prefix if you want. The input has to be text format, not .doc.  

---------------------------------------------------------------------

BBfakereads.sh

• Generating “fake” paired end reads from a single end read file

$ bfakereads.sh in=reads.fastq out1=r1.fastq out2=r2.fastq length=100

That will generate fake pairs from the input file, with whatever length you want (maximum of input read length). We use it in some cases for generating a fake LMP library for scaffolding from a set of contigs. Read 1 will be from the left end, and read 2 will be reverse-complemented and from the right end; both will retain the correct original qualities. And " /1" " /2" will be suffixed after the read name.  

------------------------------------------------------------------
Randomreads.sh

• Generate random reads

$ randomreads.sh ref=genome.fasta out=reads.fq len=100 reads=10000

"seed=-1" will use a random seed; any other value will use that specific number as the seed

You can specify paired reads, an insert size distribution, read lengths (or length ranges), and so forth. But because I developed it to benchmark mapping algorithms, it is specifically designed to give excellent control over mutations. You can specify the number of snps, insertions, deletions, and Ns per read, either exactly or probabilistically; the lengths of these events is individually customizable, the quality values can alternately be set to allow errors to be generated on the basis of quality; there's a PacBio error model; and all of the reads are annotated with their genomic origin, so you will know the correct answer when mapping.

--------------------------------------------------------------------

• Generate saturation curves to assess sequencing depth

$ bbcountunique.sh in=reads.fq out=histogram.txt

It works by pulling kmers from each input read, and testing whether it has been seen before, then storing it in a table.

The bottom line, "first", tracks whether the first kmer of the read has been seen before (independent of whether it is read 1 or read 2).

The top line, "pair", indicates whether a combined kmer from both read 1 and read 2 has been seen before. The other lines are generally safe to ignore but they track other things, like read1- or read2-specific data, and random kmers versus the first kmer.

It plots a point every X reads (configurable, default 25000).

In noncumulative mode (default), a point indicates "for the last X reads, this percentage had never been seen before". In this mode, once the line hits zero, sequencing more is not useful.

In cumulative mode, a point indicates "for all reads, this percentage had never been seen before", but still only one point is plotted per X reads.

-----------------------------------------------------------------
CalcTrueQuality.sh

http://seqanswers.com/forums/showthread.php?p=170904

In light of the quality-score issues with the NextSeq platform, and the possibility of future Illumina platforms (HiSeq 3000 and 4000) also using quantized quality scores, I developed it for recalibrating the scores to ensure accuracy and restore the full range of values.

-----------------------------------------------------------------

BBMapskimmer.sh

BBMap is designed to find the best mapping, and heuristics will cause it to ignore mappings that are valid but substantially worse. Therefore, I made a different version of it, BBMapSkimmer, which is designed to find all of the mappings above a certain threshold. The shellscript is bbmapskimmer.sh and the usage is similar to bbmap.sh or mapPacBio.sh. For primers, which I assume will be short, you may wish to use a lower than default K of, say, 10 or 11, and add the "slow" flag.

--------------------------------------------------------------

msa.sh and curprimers.sh

Quoted from Brian's response directly.

I also wrote another pair of programs specifically for working with primer pairs, msa.sh and cutprimers.sh. msa.sh will forcibly align a primer sequence (or a set of primer sequences) against a set of reference sequences to find the single best matching location per reference sequence - in other words, if you have 3 primers and 100 ref sequences, it will output a sam file with exactly 100 alignments - one per ref sequence, using the primer sequence that matched best. Of course you can also just run it with 1 primer sequence.

So you run msa twice - once for the left primer, and once for the right primer - and generate 2 sam files. Then you feed those into cutprimers.sh, which will create a new fasta file containing the sequence between the primers, for each reference sequence. We used these programs to synthetically cut V4 out of full-length 16S sequences.

I should say, though, that the primer sites identified are based on the normal BBMap scoring, which is not necessarily the same as where the primers would bind naturally, though with highly conserved regions there should be no difference.

------------------------------------------------------
testformat.sh

Identify type of Q-score encoding in sequence files

$ testformat.sh in=seq.fq.gz
sanger    fastq    gz    interleaved    150bp

--------------------------------------------------
kcompress.sh

Newest member of BBTools. Identify constituent k-mers. 
http://seqanswers.com/forums/showthread.php?t=63258

----------------------------------------------------
commonkmers.sh

Find all k-mers for a given sequence.

$ commonkmers.sh in=reads.fq out=kmers.txt k=4 count=t display=999

Will produce output that looks like

MISEQ05:239:000000000-A74HF:1:2110:14788:23085	ATGA=8	ATGC=6	GTCA=6	AAAT=5	AAGC=5	AATG=5	AGCA=5	ATAA=5	ATTA=5	CAAA=5	CATA=5	CATC=5	CTGC=5	AACC=4	AACG=4	AAGA=4	ACAT=4	ACCA=4	AGAA=4	ATCA=4	ATGG=4	CAAG=4	CCAA=4	CCTC=4	CTCA=4	CTGA=4	CTTC=4	GAGC=4	GGTA=4	GTAA=4	GTTA=4	AAAA=3	AAAC=3	AAGT=3	ACCG=3	ACGG=3	ACTG=3	AGAT=3	AGCT=3	AGGA=3	AGTA=3	AGTC=3	CAGC=3	CATG=3	CGAG=3	CGGA=3	CGTC=3	CTAA=3	CTCC=3	CTTA=3	GAAA=3	GACA=3	GACC=3	GAGA=3	GCAA=3	GGAC=3	TCAA=3	TGCA=3	AAAG=2	AACA=2	AATA=2	AATC=2	ACAA=2	ACCC=2	ACCT=2	ACGA=2	ACGC=2	AGAC=2	AGCG=2	AGGC=2	CAAC=2	CAGG=2	CCGC=2	GCCA=2	GCTA=2	GGAA=2	GGCA=2	TAAA=2	TAGA=2	TCCA=2	TGAA=2	AAGG=1	AATT=1	ACGT=1	AGAG=1	AGCC=1	AGGG=1	ATAC=1	ATAG=1	ATTG=1	CACA=1	CACG=1	CAGA=1	CCAC=1	CCCA=1	CCGA=1	CCTA=1	CGAC=1	CGCA=1	CGCC=1	CGCG=1	CGTA=1	CTAC=1	GAAC=1	GCGA=1	GCGC=1	GTAC=1	GTGA=1	TTAA=1

-----------------------------------------------------
Mutate.sh

Simulate multiple mutants from a known reference (e.g. E. coli).

$ mutate.sh in=e_coli.fasta out=mutant.fasta id=99 
$ randomreads.sh ref=mutant.fasta out=reads.fq.gz reads=5m length=150 paired adderrors

That will create a mutant version of E.coli with 99% identity to the original, and then generate 5 million simulated read pairs from the new genome. You can repeat this multiple times; each mutant will be different.

------------------------------------

Partition.sh

One can partition a large dataset with partition.sh into smaller subsets (example below splits data into 8 chunks).

partition.sh in=r1.fq in2=r2.fq out=r1_part%.fq out2=r2_part%.fq ways=8

-----------------------------------
clumpify.sh

If you are concerned about file size and want the files to be as small as possible, give Clumpify a try. It can reduce filesize by around 30% losslessly by reordering the reads. I've found that this also typically accelerates subsequent analysis pipelines by a similar factor (up to 30%). Usage:

clumpify.sh in=reads.fastq.gz out=clumped.fastq.gz

clumpify.sh in1=reads_R1.fastq.gz in2=reads_R2.fastq.gz out1=clumped_R1.fastq.gz out2=clumped_R2.fastq.gz

• Clumpify.sh can now mark/remove sequence duplicates (optical/PCR/otherwise) from NGS data

This does NOT require alignments so it should prove more useful compared to Picard MarkDuplicates. Relevant options for clumpify.sh command are listed below.

dedupe=f optical=f (default)
Nothing happens with regards to duplicates.

dedupe=t optical=f
All duplicates are detected, whether optical or not.  All copies except one are removed for each duplicate.

dedupe=f optical=t
Nothing happens.

dedupe=t optical=t

Only optical duplicates (those with an X or Y coordinate within dist) are detected.  All copies except one are removed for each duplicate.
The allduplicates flag makes all copies of duplicates removed, rather than leaving a single copy.  But like optical, it has no effect unless dedupe=t.

Note: If you set "dupedist" to anything greater than 0, "optical" gets enabled automatically.

-------------------------------------
fuse.sh

Fuse will automatically reverse-complement read 2. Pad (N) amount can be adjusted as necessary. This will for example create a full size amplicon that can be used for alignments.

fuse.sh in1=r1.fq in2=r2.fq pad=130 out=fused.fq fusepairs

Address of the bookmark: http://pacb.com/blog/the-hifi-difference-true-long-reads-vs-synthetic-long-reads/

Before you go down that path, consider this critical biological question:
Are you sequencing human cells—or bacterial contamination?

The Silent Saboteur: Mycoplasma in Cell Cultures

Mycoplasma contamination remains one of the most widespread and underdiagnosed issues in tissue culture work. Studies suggest that 15–35% of cell lines in use may be contaminated, often without visible signs. Unlike other microbial infections, Mycoplasma does not produce cloudiness, odor, or a change in pH. Many researchers won’t detect it unless they specifically test for it.

The consequences, however, are profound. Mycoplasma can significantly alter:

• Host gene expression patterns

• Cell proliferation rates

• Epigenetic profiles and chromatin accessibility

• Cytokine signaling and immune responses

In short, it can skew your results, compromise your biological conclusions, and invalidate weeks or months of research.

A Simple Diagnostic Step: Map Against Mycoplasma Genomes

If you encounter poor alignment rates to the human genome, consider mapping your reads to a Mycoplasma reference genome—or better yet, use a combined human + Mycoplasma reference. There have been cases where over half of all reads, initially assumed to be from human cells, were in fact bacterial in origin. This check is fast, easy, and could save your project.

How Contamination Happens—and Persists

Mycoplasma is small (0.1–0.3 μm), lacks a cell wall, and can pass through standard filters undetected. Common sources include:

• Contaminated reagents (e.g., FBS)

• Infected cell lines obtained from other labs

• Poor aseptic technique or shared equipment

Once present, it spreads quickly between cultures and can persist for months, silently affecting results.

Why Treatment Is Difficult

While antibiotics such as Plasmocin or BM-Cyclin are sometimes used, they often offer only partial resolution and may themselves alter cell behavior. In many cases, the best course of action is to discard the contaminated culture and start with a fresh, verified stock.

Practical Recommendations for Researchers

• Routinely test for Mycoplasma using PCR, qPCR, or fluorescence-based assays

• Incorporate contamination screens into your sequencing QC pipeline

• Use combined reference genomes when mapping ambiguous reads

• Practice strict aseptic technique and monitor all incoming cell lines

• Don’t ignore unexplained data anomalies—they might point to contamination

Closing Thought: Contamination Is a Biological Variable

It’s easy to view poor mapping as a technical issue, but sometimes the problem lies deeper—in the biology itself. Mycoplasma contamination doesn’t just interfere with sequencing; it interferes with science. As a research community, we must treat contamination not as an afterthought, but as a key variable to control.

So next time your reads won’t align, don’t just tune the aligner. Ask if your cells are telling the truth—or if they're hiding something.