Skip to content

Latest commit

 

History

History
175 lines (143 loc) · 5.87 KB

evaluation.md

File metadata and controls

175 lines (143 loc) · 5.87 KB
Raw

Assembling subreads after processing through SequelTools

The raw data was downloaded from the PacBio ftp site, that provides public accessible PacBio Sequel II dataset for many genomes. We used Arabidopsis thaliana dataset for assembly.

wget https://downloads.pacbcloud.com/public/SequelData/ArabidopsisDemoData/SequenceData/3_C01_customer/m54113_160914_092411.scraps.bam
wget https://downloads.pacbcloud.com/public/SequelData/ArabidopsisDemoData/SequenceData/3_C01_customer/m54113_160914_092411.subreads.bam
wget https://downloads.pacbcloud.com/public/SequelData/ArabidopsisDemoData/SequenceData/1_A01_customer/m54113_160913_184949.scraps.bam
wget https://downloads.pacbcloud.com/public/SequelData/ArabidopsisDemoData/SequenceData/1_A01_customer/m54113_160913_184949.subreads.bam

List of input files were generated as follows:

ls *.scraps.bam > scraps.fofn
ls *.subreads.bam > subreads.fofn

Assembling the raw dataset without filtering

Canu was used to assemble the subreads as-is. First the reads were converted to fasta format and then a config file was created with the canu parameters.

samtools fasta --threads 16 m54113_160914_092411.subreads.bam > m54113_160914_092411.subreads.fasta
samtools fasta --threads 16 m54113_160913_184949.subreads.bam > m54113_160913_184949.subreads.fasta
cat *.subreads.fasta > subreads-raw.fa

The config file for the raw dataset was as follows:

genomeSize=135m
useGrid=true
gridEngine=slurm
gridEngineResourceOption="--time=4:00:00 -N 1 --cpus-per-task=THREADS --mem-per-cpu=MEMORY"
gridOptionsCNS="--time=8:00:00 -N 1 -p freefat --cpus-per-task=8 --mem-per-cpu=40GB"

Canu was executed as follows:

source /work/LAS/mhufford-lab/shared_dir/minconda/20181213/etc/profile.d/conda.sh
conda activate denovo_asm
tdate=$(date '+%Y%m%d')
aname=athal_raw
cfg=athal.cfg
fq="subreads-raw.fa"
canu \
   -p $aname \
   -d "canu-${tdate}" \
   -s $cfg \
   -pacbio-raw $fq

The stats were calculated using the Assemblathon Script:

Assumed genome size (Mbp)     135.00

        Number of scaffolds        555
    Total size of scaffolds  124850704
Total scaffold length as percentage of assumed genome size      92.5%
           Longest scaffold    4243121
          Shortest scaffold       2357
Number of scaffolds > 1K nt        555 100.0%
Number of scaffolds > 10K nt        528  95.1%
Number of scaffolds > 100K nt        219  39.5%
Number of scaffolds > 1M nt         30   5.4%
Number of scaffolds > 10M nt          0   0.0%
         Mean scaffold size     224956
       Median scaffold size      65852
        N50 scaffold length     699106
         L50 scaffold count         46
       NG50 scaffold length     637854
        LG50 scaffold count         54
N50 scaffold - NG50 scaffold length difference      61252
                scaffold %A      31.80
                scaffold %C      18.18
                scaffold %G      18.20
                scaffold %T      31.82
                scaffold %N       0.00
        scaffold %non-ACGTN       0.00
Number of scaffold non-ACGTN nt          0

The job summary stats (MaxRSS, runtime, CPUtime) were calculated using the standard slurm command sacct

sacct --format JobId,JobName,ReqCPUS,ReqMem,ReqNodes,Elapsed,SystemCPU,CPUTime,MaxRSS,MaxVMSize,State,Start,End -u arnstrm

Assembling the PacBio dataset after filtering for CLRs >= 5Kb

Canu was used for assembly, but the reads were first processed using SequelTools. After processing, the reads were converted to fasta.

Filtering using SequelTools:

module purge
PATH=$PATH:/ptmp/LAS/arnstrm/sequelqc
ml r-devtools samtools python
SequelTools.sh -c scraps.txt -u subreads.txt -t F -C -Z 5000 -n 16 -o runSequelT-F_len
cd runSequelT-F_len
samtools fasta --threads 16 m54113_160913_184949.subSampledSubs.sam > m54113_160913_184949.subSampledSubs.fasta
samtools fasta --threads 16 m54113_160914_092411.subSampledSubs.sam > m54113_160914_092411.subSampledSubs.fasta
cat *.subSampledSubs.fasta > subSampledSubs5kb.fa

The config file for the raw dataset was as follows:

genomeSize=135m
useGrid=true
gridEngine=slurm
gridEngineResourceOption="--time=4:00:00 -N 1 --cpus-per-task=THREADS --mem-per-cpu=MEMORY"
gridOptionsCNS="--time=8:00:00 -N 1 -p freefat --cpus-per-task=8 --mem-per-cpu=40GB"

Canu was executed as follows:

source /work/LAS/mhufford-lab/shared_dir/minconda/20181213/etc/profile.d/conda.sh
conda activate denovo_asm
tdate=$(date '+%Y%m%d')
aname=athal_l5kb
cfg=athal.cfg
fq="subSampledSubs5kb.fa"
canu \
   -p $aname \
   -d "canu-${tdate}" \
   -s $cfg \
   -pacbio-l5kb $fq

The stats were calculated using the Assemblathon Script:

Assumed genome size (Mbp)     135.00

      Number of scaffolds        659
  Total size of scaffolds  126349435
Total scaffold length as percentage of assumed genome size      93.6%
         Longest scaffold    4243141
        Shortest scaffold       6520
Number of scaffolds > 1K nt        659 100.0%
Number of scaffolds > 10K nt        649  98.5%
Number of scaffolds > 100K nt        253  38.4%
Number of scaffolds > 1M nt         24   3.6%
Number of scaffolds > 10M nt          0   0.0%
       Mean scaffold size     191729
     Median scaffold size      62543
      N50 scaffold length     576605
       L50 scaffold count         60
     NG50 scaffold length     493763
      LG50 scaffold count         68
N50 scaffold - NG50 scaffold length difference      82842
              scaffold %A      31.80
              scaffold %C      18.20
              scaffold %G      18.22
              scaffold %T      31.77
              scaffold %N       0.00
      scaffold %non-ACGTN       0.00
Number of scaffold non-ACGTN nt          0

The job summary stats (MaxRSS, runtime, CPUtime) were calculated using the standard slurm command sacct

sacct --format JobId,JobName,ReqCPUS,ReqMem,ReqNodes,Elapsed,SystemCPU,CPUTime,MaxRSS,MaxVMSize,State,Start,End -u arnstrm