view test-data/summary_pbat.txt @ 20:ff6ee551b153 draft

"planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/tools/bismark commit 2d8f9bfd501f1cd82b5cf0429c1b02b3784392ab"
author bgruening
date Fri, 04 Oct 2019 11:33:27 -0400
parents aa9bf0f29a9f
children
line wrap: on
line source

Create a temporary index with the offered files from the user. Utilizing the script: bismark_genome_preparation
Generating index with: 'bismark_genome_preparation --bowtie2 /tmp/tmpuqE7r1'
Writing bisulfite genomes out into a single MFA (multi FastA) file

Bisulfite Genome Indexer version v0.22.1 (last modified: 14 April 2019)

Step I - Prepare genome folders - completed



Total number of conversions performed:
C->T:	146875
G->A:	150504

Step II - Genome bisulfite conversions - completed


Bismark Genome Preparation - Step III: Launching the Bowtie 2 indexer
Please be aware that this process can - depending on genome size - take several hours!
Settings:
  Output files: "BS_CT.*.bt2"
  Line rate: 6 (line is 64 bytes)
  Lines per side: 1 (side is 64 bytes)
  Offset rate: 4 (one in 16)
  FTable chars: 10
  Strings: unpacked
  Max bucket size: default
  Max bucket size, sqrt multiplier: default
  Max bucket size, len divisor: 4
  Difference-cover sample period: 1024
  Endianness: little
  Actual local endianness: little
  Sanity checking: disabled
  Assertions: disabled
  Random seed: 0
  Sizeofs: void*:8, int:4, long:8, size_t:8
Input files DNA, FASTA:
  genome_mfa.CT_conversion.fa
Building a SMALL index
Reading reference sizes
  Time reading reference sizes: 00:00:00
Calculating joined length
Writing header
Reserving space for joined string
Joining reference sequences
  Time to join reference sequences: 00:00:00
bmax according to bmaxDivN setting: 189039
Using parameters --bmax 141780 --dcv 1024
  Doing ahead-of-time memory usage test
  Passed!  Constructing with these parameters: --bmax 141780 --dcv 1024
Constructing suffix-array element generator
Building DifferenceCoverSample
  Building sPrime
  Building sPrimeOrder
  V-Sorting samples
  V-Sorting samples time: 00:00:00
  Allocating rank array
  Ranking v-sort output
  Ranking v-sort output time: 00:00:00
  Invoking Larsson-Sadakane on ranks
  Invoking Larsson-Sadakane on ranks time: 00:00:00
  Sanity-checking and returning
Building samples
Reserving space for 12 sample suffixes
Generating random suffixes
QSorting 12 sample offsets, eliminating duplicates
QSorting sample offsets, eliminating duplicates time: 00:00:00
Multikey QSorting 12 samples
  (Using difference cover)
  Multikey QSorting samples time: 00:00:00
Calculating bucket sizes
Splitting and merging
  Splitting and merging time: 00:00:00
Avg bucket size: 756159 (target: 141779)
Converting suffix-array elements to index image
Allocating ftab, absorbFtab
Entering Ebwt loop
Getting block 1 of 1
  No samples; assembling all-inclusive block
  Sorting block of length 756159 for bucket 1
  (Using difference cover)
  Sorting block time: xxxx
Returning block of 756160 for bucket 1
Exited Ebwt loop
fchr[A]: 0
fchr[C]: 235897
fchr[G]: 235897
fchr[T]: 386401
fchr[$]: 756159
Exiting Ebwt::buildToDisk()
Returning from initFromVector
Wrote 4446745 bytes to primary EBWT file: BS_CT.1.bt2
Wrote 189044 bytes to secondary EBWT file: BS_CT.2.bt2
Re-opening _in1 and _in2 as input streams
Returning from Ebwt constructor
Headers:
    len: 756159
    bwtLen: 756160
    sz: 189040
    bwtSz: 189040
    lineRate: 6
    offRate: 4
    offMask: 0xfffffff0
    ftabChars: 10
    eftabLen: 20
    eftabSz: 80
    ftabLen: 1048577
    ftabSz: 4194308
    offsLen: 47260
    offsSz: 189040
    lineSz: 64
    sideSz: 64
    sideBwtSz: 48
    sideBwtLen: 192
    numSides: 3939
    numLines: 3939
    ebwtTotLen: 252096
    ebwtTotSz: 252096
    color: 0
    reverse: 0
Total time for call to driver() for forward index: xxxx
Reading reference sizes
  Time reading reference sizes: 00:00:00
Calculating joined length
Writing header
Reserving space for joined string
Joining reference sequences
  Time to join reference sequences: 00:00:00
  Time to reverse reference sequence: 00:00:00
bmax according to bmaxDivN setting: 189039
Using parameters --bmax 141780 --dcv 1024
  Doing ahead-of-time memory usage test
  Passed!  Constructing with these parameters: --bmax 141780 --dcv 1024
Constructing suffix-array element generator
Building DifferenceCoverSample
  Building sPrime
  Building sPrimeOrder
  V-Sorting samples
  V-Sorting samples time: 00:00:00
  Allocating rank array
  Ranking v-sort output
  Ranking v-sort output time: 00:00:00
  Invoking Larsson-Sadakane on ranks
  Invoking Larsson-Sadakane on ranks time: 00:00:00
  Sanity-checking and returning
Building samples
Reserving space for 12 sample suffixes
Generating random suffixes
QSorting 12 sample offsets, eliminating duplicates
QSorting sample offsets, eliminating duplicates time: 00:00:00
Multikey QSorting 12 samples
  (Using difference cover)
  Multikey QSorting samples time: 00:00:00
Calculating bucket sizes
Splitting and merging
  Splitting and merging time: 00:00:00
Avg bucket size: 756159 (target: 141779)
Converting suffix-array elements to index image
Allocating ftab, absorbFtab
Entering Ebwt loop
Getting block 1 of 1
  No samples; assembling all-inclusive block
  Sorting block of length 756159 for bucket 1
  (Using difference cover)
  Sorting block time: xxxx
Returning block of 756160 for bucket 1
Exited Ebwt loop
fchr[A]: 0
fchr[C]: 235897
fchr[G]: 235897
fchr[T]: 386401
fchr[$]: 756159
Exiting Ebwt::buildToDisk()
Returning from initFromVector
Wrote 4446745 bytes to primary EBWT file: BS_CT.rev.1.bt2
Wrote 189044 bytes to secondary EBWT file: BS_CT.rev.2.bt2
Re-opening _in1 and _in2 as input streams
Returning from Ebwt constructor
Headers:
    len: 756159
    bwtLen: 756160
    sz: 189040
    bwtSz: 189040
    lineRate: 6
    offRate: 4
    offMask: 0xfffffff0
    ftabChars: 10
    eftabLen: 20
    eftabSz: 80
    ftabLen: 1048577
    ftabSz: 4194308
    offsLen: 47260
    offsSz: 189040
    lineSz: 64
    sideSz: 64
    sideBwtSz: 48
    sideBwtLen: 192
    numSides: 3939
    numLines: 3939
    ebwtTotLen: 252096
    ebwtTotSz: 252096
    color: 0
    reverse: 1
Total time for backward call to driver() for mirror index: 00:00:01
Settings:
  Output files: "BS_GA.*.bt2"
  Line rate: 6 (line is 64 bytes)
  Lines per side: 1 (side is 64 bytes)
  Offset rate: 4 (one in 16)
  FTable chars: 10
  Strings: unpacked
  Max bucket size: default
  Max bucket size, sqrt multiplier: default
  Max bucket size, len divisor: 4
  Difference-cover sample period: 1024
  Endianness: little
  Actual local endianness: little
  Sanity checking: disabled
  Assertions: disabled
  Random seed: 0
  Sizeofs: void*:8, int:4, long:8, size_t:8
Input files DNA, FASTA:
  genome_mfa.GA_conversion.fa
Building a SMALL index
Reading reference sizes
  Time reading reference sizes: 00:00:00
Calculating joined length
Writing header
Reserving space for joined string
Joining reference sequences
  Time to join reference sequences: 00:00:00
bmax according to bmaxDivN setting: 189039
Using parameters --bmax 141780 --dcv 1024
  Doing ahead-of-time memory usage test
  Passed!  Constructing with these parameters: --bmax 141780 --dcv 1024
Constructing suffix-array element generator
Building DifferenceCoverSample
  Building sPrime
  Building sPrimeOrder
  V-Sorting samples
  V-Sorting samples time: 00:00:00
  Allocating rank array
  Ranking v-sort output
  Ranking v-sort output time: 00:00:00
  Invoking Larsson-Sadakane on ranks
  Invoking Larsson-Sadakane on ranks time: 00:00:00
  Sanity-checking and returning
Building samples
Reserving space for 12 sample suffixes
Generating random suffixes
QSorting 12 sample offsets, eliminating duplicates
QSorting sample offsets, eliminating duplicates time: 00:00:00
Multikey QSorting 12 samples
  (Using difference cover)
  Multikey QSorting samples time: 00:00:00
Calculating bucket sizes
Splitting and merging
  Splitting and merging time: 00:00:00
Avg bucket size: 756159 (target: 141779)
Converting suffix-array elements to index image
Allocating ftab, absorbFtab
Entering Ebwt loop
Getting block 1 of 1
  No samples; assembling all-inclusive block
  Sorting block of length 756159 for bucket 1
  (Using difference cover)
  Sorting block time: xxxx
Returning block of 756160 for bucket 1
Exited Ebwt loop
fchr[A]: 0
fchr[C]: 386401
fchr[G]: 533276
fchr[T]: 533276
fchr[$]: 756159
Exiting Ebwt::buildToDisk()
Returning from initFromVector
Wrote 4446745 bytes to primary EBWT file: BS_GA.1.bt2
Wrote 189044 bytes to secondary EBWT file: BS_GA.2.bt2
Re-opening _in1 and _in2 as input streams
Returning from Ebwt constructor
Headers:
    len: 756159
    bwtLen: 756160
    sz: 189040
    bwtSz: 189040
    lineRate: 6
    offRate: 4
    offMask: 0xfffffff0
    ftabChars: 10
    eftabLen: 20
    eftabSz: 80
    ftabLen: 1048577
    ftabSz: 4194308
    offsLen: 47260
    offsSz: 189040
    lineSz: 64
    sideSz: 64
    sideBwtSz: 48
    sideBwtLen: 192
    numSides: 3939
    numLines: 3939
    ebwtTotLen: 252096
    ebwtTotSz: 252096
    color: 0
    reverse: 0
Total time for call to driver() for forward index: xxxx
Reading reference sizes
  Time reading reference sizes: 00:00:00
Calculating joined length
Writing header
Reserving space for joined string
Joining reference sequences
  Time to join reference sequences: 00:00:00
  Time to reverse reference sequence: 00:00:00
bmax according to bmaxDivN setting: 189039
Using parameters --bmax 141780 --dcv 1024
  Doing ahead-of-time memory usage test
  Passed!  Constructing with these parameters: --bmax 141780 --dcv 1024
Constructing suffix-array element generator
Building DifferenceCoverSample
  Building sPrime
  Building sPrimeOrder
  V-Sorting samples
  V-Sorting samples time: 00:00:00
  Allocating rank array
  Ranking v-sort output
  Ranking v-sort output time: 00:00:00
  Invoking Larsson-Sadakane on ranks
  Invoking Larsson-Sadakane on ranks time: 00:00:00
  Sanity-checking and returning
Building samples
Reserving space for 12 sample suffixes
Generating random suffixes
QSorting 12 sample offsets, eliminating duplicates
QSorting sample offsets, eliminating duplicates time: 00:00:00
Multikey QSorting 12 samples
  (Using difference cover)
  Multikey QSorting samples time: 00:00:00
Calculating bucket sizes
Splitting and merging
  Splitting and merging time: 00:00:00
Avg bucket size: 756159 (target: 141779)
Converting suffix-array elements to index image
Allocating ftab, absorbFtab
Entering Ebwt loop
Getting block 1 of 1
  No samples; assembling all-inclusive block
  Sorting block of length 756159 for bucket 1
  (Using difference cover)
  Sorting block time: xxxx
Returning block of 756160 for bucket 1
Exited Ebwt loop
fchr[A]: 0
fchr[C]: 386401
fchr[G]: 533276
fchr[T]: 533276
fchr[$]: 756159
Exiting Ebwt::buildToDisk()
Returning from initFromVector
Wrote 4446745 bytes to primary EBWT file: BS_GA.rev.1.bt2
Wrote 189044 bytes to secondary EBWT file: BS_GA.rev.2.bt2
Re-opening _in1 and _in2 as input streams
Returning from Ebwt constructor
Headers:
    len: 756159
    bwtLen: 756160
    sz: 189040
    bwtSz: 189040
    lineRate: 6
    offRate: 4
    offMask: 0xfffffff0
    ftabChars: 10
    eftabLen: 20
    eftabSz: 80
    ftabLen: 1048577
    ftabSz: 4194308
    offsLen: 47260
    offsSz: 189040
    lineSz: 64
    sideSz: 64
    sideBwtSz: 48
    sideBwtLen: 192
    numSides: 3939
    numLines: 3939
    ebwtTotLen: 252096
    ebwtTotSz: 252096
    color: 0
    reverse: 1
Total time for backward call to driver() for mirror index: 00:00:01
Running bismark with: 'bismark --bam --temp_dir /tmp/tmpi3V3GI -o /tmp/tmpi3V3GI/results --quiet --fastq -L 20 -D 15 -R 2 --pbat --un --ambiguous /tmp/tmpuqE7r1 input_1.fq'
Bowtie 2 seems to be working fine (tested command 'bowtie2 --version' [2.3.5])
Output format is BAM (default)
Alignments will be written out in BAM format. Samtools found here: '/home/abretaud/miniconda3/envs/mulled-v1-9f2317dbfb405ed6926c55752e5c11678eee3256a6ea680d1c0f912251153030/bin/samtools'
Reference genome folder provided is /tmp/tmpuqE7r1/	(absolute path is '/tmp/tmpuqE7r1/)'
FastQ format specified

Input files to be analysed (in current folder '/tmp/tmpU_oiEI/job_working_directory/000/4/working'):
input_1.fq
Library was specified as PBAT-Seq (Post-Bisulfite Adapter Tagging), only performing alignments to the complementary strands (CTOT and CTOB)
Created output directory /tmp/tmpi3V3GI/results/!

Output will be written into the directory: /tmp/tmpi3V3GI/results/

Using temp directory: /tmp/tmpi3V3GI
Temporary files will be written into the directory: /tmp/tmpi3V3GI/
Setting parallelization to single-threaded (default)

Summary of all aligner options:	-q -L 20 -D 15 -R 2 --score-min L,0,-0.2 --ignore-quals --quiet
Current working directory is: /tmp/tmpU_oiEI/job_working_directory/000/4/working

Now reading in and storing sequence information of the genome specified in: /tmp/tmpuqE7r1/

chr chrY_JH584300_random (182347 bp)
chr chrY_JH584301_random (259875 bp)
chr chrY_JH584302_random (155838 bp)
chr chrY_JH584303_random (158099 bp)

Single-core mode: setting pid to 1

Single-end alignments will be performed
=======================================

Input file is in FastQ format
Writing a G -> A converted version of the input file input_1.fq to /tmp/tmpi3V3GI/input_1.fq_G_to_A.fastq

Created G -> A converted version of the FastQ file input_1.fq (44115 sequences in total)

Input file is input_1.fq_G_to_A.fastq (FastQ)

Now running 2 instances of Bowtie 2 against the bisulfite genome of /tmp/tmpuqE7r1/ with the specified options: -q -L 20 -D 15 -R 2 --score-min L,0,-0.2 --ignore-quals --quiet

Now starting the Bowtie 2 aligner for GAreadCTgenome (reading in sequences from /tmp/tmpi3V3GI/input_1.fq_G_to_A.fastq with options -q -L 20 -D 15 -R 2 --score-min L,0,-0.2 --ignore-quals --quiet --nofw)
Using Bowtie 2 index: /tmp/tmpuqE7r1/Bisulfite_Genome/CT_conversion/BS_CT

Found first alignment:	1_1	4	*	0	0	*	*	0	0	TTATATATATTAAATAAATTAATTTTTTTTATTTATATATTAAATTTTTTAATTAATTTATTAATATTTTATAAATTTTTAAATA	AAAAAEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEAEEAEEEEEE	YT:Z:UU
Now starting the Bowtie 2 aligner for GAreadGAgenome (reading in sequences from /tmp/tmpi3V3GI/input_1.fq_G_to_A.fastq with options -q -L 20 -D 15 -R 2 --score-min L,0,-0.2 --ignore-quals --quiet --norc)
Using Bowtie 2 index: /tmp/tmpuqE7r1/Bisulfite_Genome/GA_conversion/BS_GA

Found first alignment:	1_1	4	*	0	0	*	*	0	0	TTATATATATTAAATAAATTAATTTTTTTTATTTATATATTAAATTTTTTAATTAATTTATTAATATTTTATAAATTTTTAAATA	AAAAAEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEAEEAEEEEEE	YT:Z:UU

>>> Writing bisulfite mapping results to /tmp/tmpi3V3GI/results/input_1_bismark_bt2.bam <<<

Unmapped sequences will be written to /tmp/tmpi3V3GI/results/input_1.fq_unmapped_reads.fq.gz
Ambiguously mapping sequences will be written to /tmp/tmpi3V3GI/results/input_1.fq_ambiguous_reads.fq.gz

Reading in the sequence file input_1.fq
Processed 44115 sequences in total


Successfully deleted the temporary file /tmp/tmpi3V3GI/input_1.fq_G_to_A.fastq

Final Alignment report
======================
Sequences analysed in total:	44115
Number of alignments with a unique best hit from the different alignments:	13
Mapping efficiency:	0.0%

Sequences with no alignments under any condition:	44059
Sequences did not map uniquely:	43
Sequences which were discarded because genomic sequence could not be extracted:	0

Number of sequences with unique best (first) alignment came from the bowtie output:
CT/CT:	0	((converted) top strand)
CT/GA:	0	((converted) bottom strand)
GA/CT:	11	(complementary to (converted) top strand)
GA/GA:	2	(complementary to (converted) bottom strand)

Final Cytosine Methylation Report
=================================
Total number of C's analysed:	307

Total methylated C's in CpG context:	1
Total methylated C's in CHG context:	3
Total methylated C's in CHH context:	227
Total methylated C's in Unknown context:	0

Total unmethylated C's in CpG context:	1
Total unmethylated C's in CHG context:	4
Total unmethylated C's in CHH context:	71
Total unmethylated C's in Unknown context:	0

C methylated in CpG context:	50.0%
C methylated in CHG context:	42.9%
C methylated in CHH context:	76.2%
Can't determine percentage of methylated Cs in Unknown context (CN or CHN) if value was 0


Bismark completed in xxxx

====================
Bismark run complete
====================