[ilya]

BBMerge guide recommends trimming adapters before merging -- but also, in a different place, recommends providing the adapter sequences to BBMerge. Which is best?

[Shriram369]

Program ran out of memory on large dataset: Need some tips

Hi folks,

We have a shotgun metagenomic dataset (approx. 120Gbs compressed). I want to merge paired-end reads as longer reads will increase assembly performance. And I have tried it on a small subset of data and it remarkably increased N50 and scaffold length.

But now I want to merged approx 120Gbs of compressed data for subsequent assembly. We have a system with 32 threads and 120Gb of memory. After going through tips on bbtools page, I tried following command and ran out of memory (Error message: This program ran out of memory.
Try increasing the -Xmx flag and using tool-specific memory-related parameters).

bbmerge-auto.sh in1=in_R1.fastq.gz in2=in_R2.fastq.gz out=merged.fastq.gz outu1=1_um.fastq.gz outu2=2_um.fastq.gz outa=adapters.txt ihist=insert_histogram.txt k=62 vstrict rem extend2=50 ecct mininsert=150 -Xmx80g minprob=0.8 prefilter=2 prealloc ziplevel=5

My question are:

1. Are there any other specific parameters with which it is manageable to run this command on mentioned configured server.

2. Can I subset the data using partition.sh bbtools wrapper and run the command? But as I understand sub-setting the data will reduced merging of reads. is it true?

Any tips/advice in this case is appreciated.

Thanks

[GenoMax]

@Shriram369: As long as your reads are in proper order in the files it would be fine to sub-set the data into manageable chunks and then do the merging.

[yy273826987]

Can we merge two forward reads with this tool?

Hi Brain,

I am really new to bioinformatics data analysis and just found this wonderful tool. Here I have a question: I have several environmental samples (A, B, and C). I sequenced them (shotgun metagenoimcs sequencing; paired-end) and found that, for sample B, the sequencing depth is not high enough. So, I asked the sequencing center to sequence sample B again. In the end, I got two sets of sequencing results for sample B: B.R1, B.R2, B.2nd.R1, and B.2nd.R2. For my downstream analysis (e.g., co-assembly), do you think I should merge B.R1 and B.2nd.R1 first? If so, how can use BBmerge to do that? Based on my understanding, BBmerger is designed to merge R1 and R2. Can it be used to merge two sets of R1s (from two separate sequencing runs)? Or, is that merging even necessary?

Thanks a lot!

Yours,

[GenoMax]

If you have two separate sequencing runs you can't "merge" the two reads since they are not sequencing the same fragment. Reason you can (in some cases) merge two reads R1/R2 to get a longer representation is because they are sequences from same fragment.

[Emily Shrimpton]

Hi Brian,

I'm trying to use your BBMerge program on my trimmed miRNA PE reads, but I am getting a very low merge rate. I looked at the files that had sequences unable to merge to try to understand what the problem could be, but I'm confused because there were sequences that match and could have been merged. (Please refer to the below comparison of the R1 and R2 sequences from the unmergeable files.) Could you provide some insight as to why this might be happening?

[login001: ~]$ head mirna4Merged/14343_003_R1_fastx_trimmer_NOT_MERGED_output.fastq
@A00672:72:HNTG5DSX2:4:1101:24198:13369 1:N:0:AAGTACAG
TTCAAGTAATCCAGGATAGGC
+
FFFFFFFFFFFFFFFFFFFFF
@A00672:72:HNTG5DSX2:4:1101:24795:13369 1:N:0:AAGTACAG
TGAGGTAGTAGGTTGTGTGGTTT
+
FFFFFFFFFFFFFFFFFFFFFFF
@A00672:72:HNTG5DSX2:4:1101:29351:13369 1:N:0:AAGTACAG
TATTGCACTCGTCCCGGCCTCC
[login001: ~]$
[login001: ~]$
[login001: ~]$
[login001: ~]$ head mirna4Merged/14343_003_R2_fastx_trimmer_NOT_MERGED_output.fastq
@A00672:72:HNTG5DSX2:4:1101:24198:13369 2:N:0:AAGTACAG
GCCTATCCTGGATTACTTGAA
+
FFFFFFFFFFFFFFFFFFFFF
@A00672:72:HNTG5DSX2:4:1101:24795:13369 2:N:0:AAGTACAG
AAACCACACAACCTACTACCTCA
+
FFFFFFFFFFFFFFFFFFFFFFF
@A00672:72:HNTG5DSX2:4:1101:29351:13369 2:N:0:AAGTACAG
GGAGGCCGGGACGAGTGCAATA

Thank you for your time!
Emily Shrimpton

Edit: RESOLVED

[mcoyne]

Hi all --

I'm having trouble with BBMerge, from 38.90...

Given two fastq reads, adapter trimmed, pre-merge:

r1:
@M08540:24:000000000-KVP7W:1:1101:14796:2333 1:N:0:AGCGATAG+TAATCTTA
TACTTTGCGAGATGCCCTAAGCTGGCGGGACTCTGGGGTTCGCGACACTGGCAGAGCATTACGCCCTGCAGGTAATACGACTCACTATAGGGGATAGATGTCCACGAggtctctATCATGCGGCTTTTAACAATCGTACGCTGCAGGTCGACAGATCGGAAGAGCACACGTCTGAACTCCAGTCACAGCGATAG
+
CCCCCFFFCCCCGGGGGGGGGGHHHGGGGGGHHHHHHGEEHGGGGGGGFHHHHGGHHHHHHHGGGGGHHHHHHHHHHHHFEGGHHHHHHHHHGGGHHHHHHHHHHGGGGGHHHHHHHHHHHGGGGGHHHHHHHHHGHHHGGGGGHHHHHGGGGGGGGGGGGFGGGGGGGGGGFFFFFFFFFFFFFFFFFFFFFF

r2:
@M08540:24:000000000-KVP7W:1:1101:14796:2333 2:N:0:AGCGATAG+TAATCTTA
GTCGACCTGCAGCGTACGATTGTTAAAAGCCGCATGATAGAGACCTCGTGGACATCTATCCCCTATAGTGAGTCGTATTACCTGCAGGGCGTAATGCTCTGCCAGTGTCGCGAACCCCAGAGTCCCGCCAGCTTAGGGCATCTCGCAAAGTAAGATCGGAAGAGCGTCGTGTAGGGAAAGAGTGTTAAGATTAGTGTAGATCTCGGGGGTAGCTGGAGCATTAAAAAAGAAAAAGAGATGAGAGTAGAAG
+
BBBB@DBBFFFFGGGGGGCGGGHHHHHHGHHGGGGGHHGHHHHHGHHGHHGGGHHHHHHHHHGHGHHGHHHHHHGGHGHHHGHHHHHHGGGGGGGHHGHHHHHHHHHHHGHGFGGGGGGGHHHHHHGGGGGGHHHHHHHGHHHHHGGGGHGFHHHHHHGGGGHHHGGGGGG.BCFGC:9CCFCFFGGGGFFFGGB0CBFB00;/;.@@D-.;//;.9.9:9//;///-.::/.....9/////::;////

The relationship between these two reads is (with r2 reverse-complemented, common region in lower case):

r1:                                                                                                         tactttgcgagatgccctaagctggcgggactctggggttcgcgacactggcagagcattacgccctgcaggtaatacgactcactataggggatagatgtccacgaggtctctatcatgcggcttttaacaatcgtacgctgcaggtcgacAGATCGGAAGAGCACACGTCTGAACTCCAGTCACAGCGATAG
r2(rc):   CTTCTACTCTCATCTCTTTTTCTTTTTTAATGCTCCAGCTACCCCCGAGATCTACACTAATCTTAACACTCTTTCCCTACACGACGCTCTTCCGATCTtactttgcgagatgccctaagctggcgggactctggggttcgcgacactggcagagcattacgccctgcaggtaatacgactcactataggggatagatgtccacgaggtctctatcatgcggcttttaacaatcgtacgctgcaggtcgac

My desired output is the merged read with all the data included, with no trimming on either end, i.e.:

CTTCTACTCTCATCTCTTTTTCTTTTTTAATGCTCCAGCTACCCCCGAGATCTACACTAATCTTAACACTCTTTCCCTACACGACGCTCTTCCGATCTtactttgcgagatgccctaagctggcgggactctggggttcgcgacactggcagagcattacgccctgcaggtaatacgactcactataggggatagatgtccacgaggtctctatcatgcggcttttaacaatcgtacgctgcaggtcgacAGATCGGAAGAGCACACGTCTGAACTCCAGTCACAGCGATAG

BBMerge switches used (additional file input and output options are also included):

qtrim=f qtrim2=f trimq=0 tbo=f tno=f usequality=f forcetrimleft=0 forcetrimright=0 forcetrimright2=0 forcetrimmod=0 forcemerge=t

produces this merged output:

>M08540:24:000000000-KVP7W:1:1101:14796:2333 1:N:0:AGCGATAG+TAATCTTA
TACTTTGCGAGATGCCCTAAGCTGGCGGGACTCTGGGGTTCGCGACACTGGCAGAGCATTACGCCCTGCA
GGTAATACGACTCACTATAGGGGATAGATGTCCACGAGGTCTCTATCATGCGGCTTTTAACAATCGTACG
CTGCAGGTCGAC

In other words, it appears to be doing "trimbyoverlap" (tbo) even though it's explicitly set to off.

Are there setting I can use that will produce the desired output (a 292 bp sequence, with no trimming on either end)?

[kubu4]

Hi Brian,

I'm trying to use your BBMerge program on my trimmed miRNA PE reads, but I am getting a very low merge rate. I looked at the files that had sequences unable to merge to try to understand what the problem could be, but I'm confused because there were sequences that match and could have been merged. (Please refer to the below comparison of the R1 and R2 sequences from the unmergeable files.) Could you provide some insight as to why this might be happening?

[login001: ~]$ head mirna4Merged/14343_003_R1_fastx_trimmer_NOT_MERGED_output.fastq
@A00672:72:HNTG5DSX2:4:1101:24198:13369 1:N:0:AAGTACAG
TTCAAGTAATCCAGGATAGGC
+
FFFFFFFFFFFFFFFFFFFFF
@A00672:72:HNTG5DSX2:4:1101:24795:13369 1:N:0:AAGTACAG
TGAGGTAGTAGGTTGTGTGGTTT
+
FFFFFFFFFFFFFFFFFFFFFFF
@A00672:72:HNTG5DSX2:4:1101:29351:13369 1:N:0:AAGTACAG
TATTGCACTCGTCCCGGCCTCC
[login001: ~]$
[login001: ~]$
[login001: ~]$
[login001: ~]$ head mirna4Merged/14343_003_R2_fastx_trimmer_NOT_MERGED_output.fastq
@A00672:72:HNTG5DSX2:4:1101:24198:13369 2:N:0:AAGTACAG
GCCTATCCTGGATTACTTGAA
+
FFFFFFFFFFFFFFFFFFFFF
@A00672:72:HNTG5DSX2:4:1101:24795:13369 2:N:0:AAGTACAG
AAACCACACAACCTACTACCTCA
+
FFFFFFFFFFFFFFFFFFFFFFF
@A00672:72:HNTG5DSX2:4:1101:29351:13369 2:N:0:AAGTACAG
GGAGGCCGGGACGAGTGCAATA

Thank you for your time!
Emily Shrimpton

Edit: RESOLVED

Emily Shrimpton - Would you be willing to post how you resolved this (if you still remember)? Thanks!

[ghimbikal]

Hi, I am new to programming and bioinformatics. I am having trouble merging my fastq files with the inclusion of the overhangs/nonoverlapping reads at either end. I tried PEAR before, which gave me final consensus only for the overlapping region. I tried bbmerge and it gave me the same thing. And I am not able to figure out what script I should use to have the final consensus that includes overhangs added to ends of my consensus so I have the full consensus sequences. For example, I have following fastq reads. Could any expert help me with a custom script for this? thanks.

@S1_A01_015_F
ATTTATTTTTGGTGCTTTTTCTGGTGTAGTAGGAACTACATTATCTGTTTTAATTAGAATGGAATTAGCACAACCCGGTAATCAAATTTTTGCTGGGAATCATCATTTATATAATGTTGTTGTTACAGCACATGCATTTATTATGATTTTTTTTATGGTTATGCCTGTTTTAATAGGTGGTTTTGGTAATTGGTTTGTACCTTTAATGATTGGTGCTCCAGATATGGCTTTTCCTCGTATGAATAATATAAGTTTTTGGTTATTACCACCATCATTATTATTATTAGTTTCTTCAGCTATTGTTGAATCAGGTGCAGGTACTGGTTGGACTGTATATCCTCCTTTATCAAGTGTACAAGCACATTCAGGTCCTTCAGTAGATTTAGCTATTTTTAGTTTACATTTATCAGGTATTTCTTCTTTATTAGGTGCTATTAATTTTTTATCTACTATTTATAATATGAGAGCTCCAGGTTTAAGTTTTCATAGATTACCTTTATTTGTTTGGGCTATATTTATTACTGCTTTTTTATTATTATTAACTTTACCTGTATTAGCTGGTGCAATTACTATGTTATTAACTGATAGAAACTTAAATACATCTTTTTACGATCCATCAGGCGGAGGAGATCCTGTATTATACCAACATTTATTTTGGTTTTTCGGCAACCCCGGAAG
+
9>*%**ROOAB*,78K[[[[W>W:0G6J@RP_J__Y_TPK_W_MRP\\__\\__\\\_W;W___\_\W___\__________\___W____\\_W____\___\_________WK__RW____\__\__________RKWWW__RW__W_WRW________________WWWW_\YW___WWWW______WW_WWW_____W_________________\S______LRRW_____Y____________\_W_________________________________\__________________W________________________________WRRWWW___________________________________________________\_____________________________________________RRR_____________________O_____________________________________________________[[[[R[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[WW[[[[[[[[[[[[[[[[[[[[[R[[[[[[W[T[[[[[[[[[[[[R[[[[[[W[[R[[[MQ[[SO[W[LJB[[HMLWRT[[UPRSSSSJ=ISLM[ISIHKSS>@[IDD''&&&&*<A5-*((5

@S1_A01_015_R
CTCCTCGCCTGATGGATCGTAAAAGATGTATTTAAGTTTCTATCAGTTAATAACATAGTAATTGCACCAGCTAATACAGGTAAAGTTAATAATAATAAAAAAGCAGTAATAAATATAGCCCAAACAAATAAAGGTAATCTATGAAAACTTAAACCTGGAGCTCTCATATTATAAATAGTAGATAAAAAATTAATAGCACCTAATAAAGAAGAAATACCTGATAAATGTAAACTAAAAATAGCTAAATCTACTGAAGGACCTGAATGTGCTTGTACACTTGATAAAGGAGGATATACAGTCCAACCAGTACCTGCACCTGATTCAACAATAGCTGAAGAAACTAATAATAATAATGATGGTGGTAATAACCAAAAACTTATATTATTCATACGAGGAAAAGCCATATCTGGAGCACCAATCATTAAAGGTACAAACCAATTACCAAAACCACCTATTAAAACAGGCATAACCATAAAAAAAATCATAATAAATGCATGTGCTGTAACAACAACATTATATAAATGATGATTCCCAGCAAAAATTTGATTACCGGGTTGTGCTAATTCCATTCTAATTAAAACAGATAATGTAGTTCCTACTACACCAGAAAAAGCACCAAAAATTAAATATAAAGTACCTATGTCTTTATGATTTGTTGAAAA
+
B*+'2+(0C2W:0[C4'L4(*'*)/=1:?HJ[J<=7____HH_GJI_FS_Y__PP_R___TN_\\T_\M\_\_\\_\OY__\_\_______RR__\_________\__W______\OSYYW_Y___\\\\YCLW_____W_\WW\______\\\\Y_YWYWW___Y\_______W_\SAG<KQY_WRW_____YWQWY\_O\\EYYQN_WC_\Y___\Y_\\__\___\\\_\_\_____Y__\_\__\__YWW_Y_\__\_WQY_Y_Y___\\_\_Y__KYYWLLW_SWWO\___OY\_\Y____\_______\____\__________YYLLWQE_\\___R_____\\\YMYYOOY___________YYYYY\\OOY\\\________N_SRFG__Y\\YYY_YORR_Y__Y__YR___YYYOY__________Y____LYY_OO_Y_____L_Y___YY_________O______________YYLJOYH?=OL___________MU_U___________[NUNMOUURWOQUUUPIIW[OMWOU7?OUUUPWU[[[UUN[WNLW[[WWUWNMU[[U[U[UDLC[QSHRWRSPWQMPQS[VTS[JJPPS[TTTT[[[PQR[[[[[RRJHMRTQQRR[W[R[[LQQSW@MLA9J1QQKL

could any expert provide me a custom script for this two fastq files to merge so i get full lenght consensus?

I used the command
# Construct the BBMerge command with options to not trim the overhangs
bbmerge_cmd = f'bbmerge.sh in1={forward_file} in2={reverse_file} out={output_file} outu={output_file}_unmerged.fastq ' \
f'qtrim=f qtrim2=f trimq=0 tbo=f tno=f usequality=f forcetrimleft=0 forcetrimright=0 forcetrimright2=0 forcetrimmod=0'

But still it gave me output by trimming the overlapped region as output as below:
@S1_A01_015
TTTAATTTTTGGTGCTTTTTCTGGTGTAGTAGGAACTACATTATCTGTTTTAATTAGAATGGAATTAGCACAACCCGGTAATCAAATTTTTGCTGGGAATCATCATTTATATAATGTTGTTGTTACAGCACATGCATTTATTATGATTTTTTTTATGGTTATGCCTGTTTTAATAGGTGGTTTTGGTAATTGGTTTGTACCTTTAATGATTGGTGCTCCAGATATGGCTTTTCCTCGTATGAATAATATAAGTTTTTGGTTATTACCACCATCATTATTATTATTAGTTTCTTCAGCTATTGTTGAATCAGGTGCAGGTACTGGTTGGACTGTATATCCTCCTTTATCAAGTGTACAAGCACATTCAGGTCCTTCAGTAGATTTAGCTATTTTTAGTTTACATTTATCAGGTATTTCTTCTTTATTAGGTGCTATTAATTTTTTATCTACTATTTATAATATGAGAGCTCCAGGTTTAAGTTTTCATAGATTACCTTTATTTGTTTGGGCTATATTTATTACTGCTTTTTTATTATTATTAACTTTACCTGTATTAGCTGGTGCAATTACTATGTTATTAACTGATAGAAACTTAAATACATCTTTTTANGATCCANCAGGCGGAGG

But i wanted the consensus as

TTTTCAACAAATCATAAAGACATAGGTACTTTATATTTAATTTTTGGTGCTTTTTCTGGTGTAGTAGGAACTACATTATCTGTTTTAATTAGAATGGAATTAGCACAACCCGGTAATCAAATTTTTGCTGGGAATCATCATTTATATAATGTTGTTGTTACAGCACATGCATTTATTATGATTTTTTTTATGGTTATGCCTGTTTTAATAGGTGGTTTTGGTAATTGGTTTGTACCTTTAATGATTGGTGCTCCAGATATGGCTTTTCCTCGTATGAATAATATAAGTTTTTGGTTATTACCACCATCATTATTATTATTAGTTTCTTCAGCTATTGTTGAATCAGGTGCAGGTACTGGTTGGACTGTATATCCTCCTTTATCAAGTGTACAAGCACATTCAGGTCCTTCAGTAGATTTAGCTATTTTTAGTTTACATTTATCAGGTATTTCTTCTTTATTAGGTGCTATTAATTTTTTATCTACTATTTATAATATGAGAGCTCCAGGTTTAAGTTTTCATAGATTACCTTTATTTGTTTGGGCTATATTTATTACTGCTTTTTTATTATTATTAACTTTACCTGTATTAGCTGGTGCAATTACTATGTTATTAACTGATAGAAACTTAAATACATCTTTTTANGATCCANCAGGCGGAGGATCCTGTATTATACCAACATTTATTTTGGTTTTTCGGCAACCCCGGAAG

[mcoyne]

I tripped over something to get around the merging problem in bbmerge described above...

Merging fastq reads with bbmerge does trim the merge, resulting in only retaining the overlapping part, however -- if I first turn the fastq reads into fasta reads and interleave them, it produces the desired output:

Checking output:

First two reads from interleaved file:

>LH00315.67.22H3LVLT3.1.1101.5558.1048.R1 1:N:0:ACTTCCGG+GACCAATT
GGGACTGTGGCCGTCGACCTGCAGCGTACGGGAATACCTGTTGATATTTAAGAGACCTCGTGGACATCTA
TCCCCTATAGTGAGTCGTATTACCTGCAGGGCGTAATG
>LH00315.67.22H3LVLT3.1.1101.5558.1048.R2 2:N:0:ACTTCCGG+GACCAATT
TCGGACCAACTAAGCTGGCGGGACTCTGGGGTTCGCGACACTGGCAGAGCATTACGCCCTGCAGGTAATA
CGACTCACTATAGGGGATAGATGTCCACGAGGTCTCTT


>2nd read RC
AAGAGACCTCGTGGACATCTATCCCCTATAGTGAGTCGTATTACCTGCAGGGCGTAATGCTCTGCCAGTGT
CGCGAACCCCAGAGTCCCGCCAGCTTAGTTGGTCCGA

Blast alignment:

Query  50 AAGAGACCTCGTGGACATCTATCCCCTATAGTGAGTCGTATTACCTGCAGGGCGTAATG 108
          |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Sbjct 108 AAGAGACCTCGTGGACATCTATCCCCTATAGTGAGTCGTATTACCTGCAGGGCGTAATG 50

Manual alignment:

GGGACTGTGGCCGTCGACCTGCAGCGTACGGGAATACCTGTTGATATTTaagagacctcgtggacatctatcccctatagtgagtcgtattacctgcagggcgtaatg
                                                 aagagacctcgtggacatctatcccctatagtgagtcgtattacctgcagggcgtaatgCTCTGCCAGTGTCGCGAACCCCAGAGTCCCGCCAGCTTAGTTGGTCCGA

Expected sequence (157 bp):
GGGACTGTGGCCGTCGACCTGCAGCGTACGGGAATACCTGTTGATATTTaagagacctcgtggacatctatcccctatagtgagtcgtattacctgcagggcgtaatgCTCTGCCAGTGTCGCGAACCCCAGAGTCCCGCCAGCTTAGTTGGTCCGA

BBMerge output (157 bp):

>LH00315.67.22H3LVLT3.1.1101.5558.1048.R1 1:N:0:ACTTCCGG+GACCAATT
GGGACTGTGGCCGTCGACCTGCAGCGTACGGGAATACCTGTTGATATTTAAGAGACCTCGTGGACATCTA
TCCCCTATAGTGAGTCGTATTACCTGCAGGGCGTAATGCTCTGCCAGTGTCGCGAACCCCAGAGTCCCGC
CAGCTTAGTTGGTCCGA

Not sure if combining the R1 and R2 fastq files into a single interleaved file before running bbmerge, or if producing two fasta files rather than one interleaved fasta file will have the same effect.