[csmiller]

usejni and compiled C code in BBTools

I just installed the latest version of the BBTools (38.26), and I notice that the C code provided by the usejni=t flag for some tools has been depreciated / disabled.

I found this in the changelog:

Removed JNI path flag from BBMerge, BBMap, and RQCFilter shell scripts.

and this in docs/compiling.txt:

3) C code. This was developed by Jonathan Rood to accelerate BBMap, BBMerge, and Dedupe, but is currently disabled.

Sure enough, it is commented out in the bbmap.sh code:

        #local CMD="java -Djava.library.path=$NATIVELIBDIR $EA $z -cp $CP align2.BBMap build=1 overwrite=true fastareadlen=500 $@"
        local CMD="java $EA $z -cp $CP align2.BBMap build=1 overwrite=true fastareadlen=500 $@"

If I revert to the previous version of the CMD, with the java.library.path set, then the command runs with the compiled C code just fine.

Why was this disabled? Does this affect previous analyses that used this C code? That is, does the C code contain an error that means usejni=t in previous versions will produce different output than the java-only code? Or was this purely a performance or compatibility issue, or something else?

Sorry if I've missed this already posted somewhere, and thanks in advance for any help.

Chris

[csmiller]

How to use usejni with latest version (

I just installed the latest version of the BBTools (38.26), and I can't seem to get the usejni flag to work. The java .so file compiles fine, but then I get error messages like this when I run, e.g., bbmap.sh with usejni=t:

Native library can not be found in java.library.path.

I found this in the changelog:

Removed JNI path flag from BBMerge, BBMap, and RQCFilter shell scripts.

and this in docs/compiling.txt:

3) C code. This was developed by Jonathan Rood to accelerate BBMap, BBMerge, and Dedupe, but is currently disabled.

And sure enough, it is commented out in the bbmap.sh code:

        #local CMD="java -Djava.library.path=$NATIVELIBDIR $EA $z -cp $CP align2.BBMap build=1 overwrite=true fastareadlen=500 $@"
        local CMD="java $EA $z -cp $CP align2.BBMap build=1 overwrite=true fastareadlen=500 $@"

If I revert to the previous version of the CMD, with the java.library.path set, then the command runs with the C code fine.

Why was this disabled? Does this affect previous analyses that used this C code? Or was this purely a performance issue?

Sorry if I've missed this posted somewhere else, and thanks in advance for any help.

Chris

[SNPsaurus]

I am sorry if this question is very basic but I am getting a low percentage of mapping reads to the reference genome, about the 36% of the pct reads mapped. Any clue what this is the case?

I am using as the reference genome the genome in scaffolds and paired-end reads...

Have you trimmed adapters away from the reads (short fragments will create reads that are part genomic and part adapter and may not map). You could use the related BBmap tool sendsketch to get a sense of what is in your reads (after trimming). When we do genotyping of samples, many samples have contaminating species...so using sendsketch can help figure out what is in there. You can input the entire fastq file with sendsketch, or go to read mose and get a result on a per read basis.

You can also grab 100 reads, turn them into fasta format and do blastn with them (if online use the blastn rather than megablast option) and see read by read what is in there.

Other options...your sample is not highly related to the reference, the reference may be incomplete and missing regions, the reference is lacking high copy repeat content like mtDNA or chloroplast and many reads go to those.

[juanita]

ref input for BBMap and paired ends

I am sorry if this question is very basic but I am getting a low percentage of mapping reads to the reference genome, about the 36% of the pct reads mapped. Any clue what this is the case?

I am using as the reference genome the genome in scaffolds and paired-end reads...

[raw937]

bbmap for demultiplexing dual barcodes.

Hello,
I need it if possible to use dual indexes.

For example: In bold dual barcode

#R1 read
@SOLEXA1_0069_FC:3:1:1673:948#ACAGTG/1
GACTAACCGGATTAGATACCCTGGTAGTCCACGCCGTAAACGATGAATGTTAGCCGTCGGGCAGTATACTGTTCGG
+
BMMQNTWSWWb_____b_bb__________Y_________YYYYY[[[Y[__________XXRWXVVVVTYYYYYT

#R2 read
@SOLEXA1_0069_FC:3:1:1673:948#ACAGTG/2
CTGAAGGGTTGCGCTCGTTGCGGGACTTAACCCAACATCTCACGACACGAGCTGACGACAGCCATGCAGCACCTGT
+
ghgaggfghhhhhhhhhhghhhhhhhhhhfhhhghfWffch[hhgahhedffddR[^W^Zc^_cac[Wb]^W^

Here are 16 possible in the file I am working on.
TCAG-TCAG
CTGA-CTGA
TCAG-GACT
GACT-GACT
AGTC-AGTC
GACT-TCAG
GACT-AGTC
GACT-CTGA
TCAG-CTGA
AGTC-TCAG
AGTC-GACT
CTGA-AGTC
CTGA-GACT
AGTC-CTGA
TCAG-AGTC
CTGA-TCAG

The first four nts are the barcode like our example before would be:
GACT-CTGA_R1.fq
GACT-CTGA_R2.fq

But you would need both reads to tell you that it's GACT-CTGA and not something else.
What would the command look like for this? Does this demux script do the dual barcoding?

[ellybelly]

bbmap aborts after mapping some reads

Hello Brian,

we are using bbmap to see in how far it is possible to quantify gene expression by mapping Illumina RNA-seq reads to the genome of a closely related species, e.g. map chimpanzee reads to human or as in this example Macaque reads.

To this end, we generated Macaque Illumina SE reads using flux-simulator and map them to
hg38 and for comparison we were also trying also Mmul8, downloaded from ensembl (wget ftp://ftp.ensembl.org/pub/release-92...toplevel.fa.gz).

Everything mapped fine to hg38, but not to Mmul8.

Exception in thread "Thread-12" java.lang.AssertionError
at align2.BBIndex.extendScore(BBIndex.java:2612)
at align2.BBIndex.slowWalk3(BBIndex.java:1389)
at align2.BBIndex.find(BBIndex.java:777)
at align2.BBIndex.find(BBIndex.java:623)
at align2.BBIndex.findAdvanced(BBIndex.java:400)
at align2.AbstractMapThread.quickMap(AbstractMapThread.java:750)
at align2.BBMapThread.processRead(BBMapThread.java:408)
at align2.AbstractMapThread.run(AbstractMapThread.java:508)

I tried to run on one thread, increased memory to 101G, removed small contigs of <100kb ... but the error message remains the same.

We are running a Debian system with java version "1.8.0_181" and have BBMap version 38.02 -- the detailed error output is in the attached file.

The false Mapping Rates of bbmap are so much better than for STAR & GSNAP, that we definitely want to use bbmap for our paper and we are nearly done all other species (marmoset, gorilla, chimpanzee and orangutan) and the simulations ran through -- the only missing piece is the mapping to the Mmul8.

Any help would be greatly appreciated.

Best, Ines

[JenBarb]

mkf argument in bbduk.sh (bbmap tool)

Hello,
I am trying to use the flag mkf (minkmerfraction) and I am getting an error that that argument does not exist.
sh /data/barbj/bbmap/bbduk.sh in=./../Stool_001-01.fastq outm=v2fstoolfq.fa literal=CTCAAACTTGGGTAATTAAACC k=17 mkf=0.8
java -Djava.library.path=/data/barbj/bbmap/jni/ -ea -Xmx39767m -Xms39767m -cp /data/barbj/bbmap/current/ jgi.BBDukF in=./../Stool_001-01.fastq outm=v2fstoolfq.fa literal=CTCAAACTTGGGTAATTAAACC k=17 mkf=0.8
Executing jgi.BBDukF [in=./../Stool_001-01.fastq, outm=v2fstoolfq.fa, literal=CTCAAACTTGGGTAATTAAACC, k=17, mkf=0.8]

Exception in thread "main" java.lang.RuntimeException: Unknown parameter mkf=0.8
at jgi.BBDukF.<init>(BBDukF.java:402)
any ideas why this is not working?

Jen

[Meyana]

Hi,
Hoping somebody can help me with this.

I used BBMap and now I would like to extract the reads from by .bam file that are split (/chimeric?) ie. reads that indicate a deletion.

I tried to use samblaster, but it doesn't recognize any reads as split...
(samtools view -h in.bam | samblaster -a -s split.sam -o /dev/null)
Are the split reads marked differently in BBMap compared to other aligners causing samblaster to fail?

IGV shows a good amount of reads with deletions and I can also call deletions using BBTools callvariants.sh - so I know they are in there. I just have a feeling callvariants is calling fewer deletions and with lower coverage than what IGV suggests, so I want to check up on it.

[JenBarb]

Thank you! Love the tool!

[HESmith]

@JenBarb see this thread in Biostars.

[JenBarb]

pull out sequences with matching primers

Hi Brian,
I was wondering if bbmap has a tool that will pull out reads matching a particular primer sequences? I have fastq files with amplicons from 12 different primers in the same file so i want to make subsets of the reads having specific primers of interest from this.

i have used your tool for other tasks so i figured I would ask if it also has this capability?

Thank you,
Jen

[sunnycqcn]

Hello Brian,
After running mapPacBio.sh, how can I combine the sequence of the same ID?
for example I want to combine the sequences as following:
m151006_234406_42219_c100867912550000001823195203031665_s1_p0/110457/57769_70466 id=3_0_part_2_6
m151006_234406_42219_c100867912550000001823195203031665_s1_p0/110457/57769_70466 id=3_0_part_3

Thanks,
Fuyou

[olgabot]

Add hg19 masked reference to distribution

Hello,
I'm using BBTools via bioconda and the corresponding quay.io docker container. The image has the necessary resources, e.g. the adapters fasta file:

(base) 
 Wed 25 Jul - 17:10  ~/code/tick-genome/reflow   origin ☊ master 9☀ 1● 
  docker run -it -v $PWD:/data quay.io/biocontainers/bbmap:38.06--2 bash
bash-4.2# find . -name adapters.fa
./usr/local/opt/bbmap-38.06/resources/adapters.fa
bash-4.2# cd ./usr/local/opt/bbmap-38.06/resources
bash-4.2# ll
bash: ll: command not found
bash-4.2# ls 
adapters.fa                          blacklist_silva_species_500.sketch   lambda.fa.gz                         nextera_LMP_linker.fa.gz             primes.txt.gz                        sequencing_artifacts.fa.gz
adapters_no_transposase.fa.gz        contents.txt                         lfpe.linker.fa.gz                    pJET1.2.fa                           remote_files.txt                     short.fa
blacklist_img_species_300.sketch     crelox.fa.gz                         mtst.fa                              phix174_ill.ref.fa.gz                remote_files_old.txt                 truseq.fa.gz
blacklist_nt_species_1000.sketch     favicon.ico                          nextera.fa.gz                        phix_adapters.fa.gz                  sample1.fq.gz                        truseq_rna.fa.gz
blacklist_refseq_species_250.sketch  kapatags.L40.fa                      nextera_LMP_adapter.fa.gz            polyA.fa.gz                          sample2.fq.gz

However, the removehuman.sh script uses a hardcoded path for the masked human genome posted in the RemoveHuman thread.

	local CMD="java -Djava.library.path=$NATIVELIBDIR $EA $z -cp $CP align2.BBMap minratio=0.9 maxindel=3 bwr=0.16 bw=12 quickmatch fast minhits=2 path=/global/projectb/sandbox/gaag/bbtools/hg19 pigz unpigz zl=6 qtrim=r trimq=10 untrim idtag usemodulo printunmappedcount usejni ztd=2 kfilter=25 maxsites=1 k=14 $@

Can the masked genome be included in the distribution?

Thank you!
Warmest,
Olga

[StephCarr]

You're right. It wasn't downloaded correctly. At first I used git to download the bbmap package. But when I just downloaded with wget from https://sourceforge.net/projects/bbm...p_38.12.tar.gz everything was organized correctly.

Thanks for the help.

[GenoMax]

Is your bbmap installed correctly? Have you moved any files around? I am able to run "bbfakereads.sh" and generate fastq and fasta files without a problem.