Example data retrieval command sequence:

/Users/mollydavis333/Desktop/sratoolkit.3.0.0-mac64/bin/fasterq-dump -S SRR15111745
rsync --progress /Users/mollydavis333/Desktop/SRR15111737_1.fastq  mdavi258@hpc.uncc.edu:/nobackup/tomato_genome/alt_splicing

Update:

• SRR files are downloaded into /nobackup/tomato_genome/alt_splicing
• SRR files are in two directories SRP252265 (positive control) and SRP328042 (experiment pollen tube samples)

To-do:

• Create and review samples file (used by nextflow to run the pipeline) - in progress
• Get tomato configuration file needed for nf-core/rnaseq with nextflow version 3.4 - DONE (see https://bitbucket.org/hotpollen/flavonoid-rnaseq and Ann's fork of same)
• Get (deploy) new tomato genome file from IGB Quickload - DONE
• Explain how to run nextflow - in progress
• Create GTF file for new genome models - DONE

Forked flavonoid-rnaseq into Ann's bitbucket account and cloned the fork onto cluster to: /nobackup/tomato_genome/scripts/flavonoid-rnaseq

Created a branch IGBF-3143 for this alt splicing work.

Modified "doIt.sh" script to use star/salmon alignment option.

Added documentation in the comments.

Started long-lived interactive session and loaded nf-core with module load command.

Downloaded nf-core/rnaseq v. 3.4 to: /nobackup/tomato_genome/scripts/nf-core-rnaseq-3.4 and saved singularity images to: /nobackup/tomato_genome/scripts/nxf_singularity_cachedir

Made gene model GTF format file using bed2gtf.py from https://bitbucket.org/lorainelab/genomesource/src/master/, using option "-g" to indicate that the gene name is in field 13 of input bed-detail file from http://lorainelab-quickload.scidas.org/quickload/S_lycopersicum_Jun_2022/S_lycopersicum_Jun_2022.bed.gz, revision 164 of the file, from repository browsable at https://svn.bioviz.org/viewvc/genomes/quickload/S_lycopersicum_Jun_2022/.

First run showed problem with sample sheet:

Error executing process > 'NFCORE_RNASEQ:RNASEQ:INPUT_CHECK:SAMPLESHEET_CHECK (SRP328042.csv)'

Caused by:
Process `NFCORE_RNASEQ:RNASEQ:INPUT_CHECK:SAMPLESHEET_CHECK (SRP328042.csv)` terminated with an error exit status (127)

Command executed:

check_samplesheet.py \
SRP328042.csv \
samplesheet.valid.csv

cat <<-END_VERSIONS > versions.yml
SAMPLESHEET_CHECK:
python: $(python --version | sed 's/Python //g')
END_VERSIONS

Sample sheeted "strandedness" column contains an invalid value: "stranded." Documentation mentions only three values are allowed:

Must be one of unstranded, forward or reverse.

Also, it appears that our sample files need to be "gzipped":

File has to be gzipped and have the extension ".fastq.gz" or ".fq.gz".

[~molly] and [~RobertReid] - could you "gzip" the fastq files? Also, kindly modify the "strandedness" field to one of the three options listed.

Looks like the positive control is indeed stranded, but I'm not sure whether we should "reverse" or "forward" option. I am looking into it....
attn: [~RobertReid] and [~molly]

I think we need to use "reverse" in the "strandedness" column because we actually end up sequencing the reverse (antisense) strand of the cDNA that gets made from the RNA. What happens with this protocol is that you start out by synthesizing cDNA from mRNA. The first strand synthesis step creates an anti-sense copy of the RNA, complementary to the original mRNA. That step uses regular nucleotides. The second strand synthesis step, which re-reproduces the original mRNA, using the first strand synthesis product as a template, uses dUTP in place of dTTP, which ultimately gets degraded after all the primer ligation steps happen. So, when you go to sequence the library, the only DNA you can sequence came from the first strand synthesis, which was the "reverse" strand of the original RNA. I think this is what the pipeline authors meant to intend by asking for "reverse" or "forward" in the sample sheet.

However, if this is wrong, the pipeline will let us know. It includes a step that looks at strandedness. So I think we can just go ahead and insert "reverse" in the sample sheet and see what happens.

attn: [~RobertReid] and [~molly]

Everything has been properly updated and I ran the pipeline again, making out.3.txt and err.3.txt.

However, a new error happened:

Error executing process > 'NFCORE_RNASEQ:RNASEQ:QUANTIFY_STAR_SALMON:SALMON_TXIMPORT'

Caused by:
Process `NFCORE_RNASEQ:RNASEQ:QUANTIFY_STAR_SALMON:SALMON_TXIMPORT` terminated with an error exit status (127)

Command executed:

salmon_tximport.r \
NULL \
salmon \
salmon.merged

cat <<-END_VERSIONS > versions.yml
SALMON_TXIMPORT:
r-base: $(echo $(R --version 2>&1) | sed 's/^.R version //; s/ .$//')
bioconductor-tximeta: $(Rscript -e "library(tximeta); cat(as.character(packageVersion('tximeta')))")
END_VERSIONS

Command exit status:
127

Command output:
(empty)

Command error:
INFO: Converting SIF file to temporary sandbox...
WARNING: Skipping mount /usr/local/singularity/var/singularity/mnt/session/etc/resolv.conf [files]: /etc/resolv.conf doesn't exist in container
.command.sh: line 3: salmon_tximport.r: command not found
INFO: Cleaning up image...

Work dir:
/nobackup/tomato_genome/alt_splicing/SRP328042/work/f1/9c6b783f8bafba25eb2aee017dfb65

Tip: you can replicate the issue by changing to the process work dir and entering the command `bash .command.run`

Same error occurred. Ann, Rob, and Molly met via zoom to trouble-shoot. Rob suggested trying to run the pipeline using the older configuration and older data, since that worked. Rob & Molly proceeding with trouble-shooting as I am out of ideas!

Nextflow nf-core/rnaseq pipeline update:

Directory Location: /nobackup/tomato_genome/alt_splicing/SRP252265

Dr. Reid and I successfully ran the pipeline with SRP252265 for out.4.txt:

[nf-core/rnaseq] Pipeline completed successfully
Completed at: 02-Aug-2022 16:48:38
Duration : 1m 38s
CPU hours : 481.7 (100% cached)
Succeeded : 2
Cached : 828_

Solution: Added module load singularity to steps and removed the symbolic link for nf-core-rnaseq-3.4 and directly downloaded the files to the directory while not in an interactive session:

module load nf-core
nf-core download rnaseq -r 3.4

The only symbolic links present were for the file resources:

S_lycopersicum_Jun_2022.bed
S_lycopersicum_Jun_2022.fa
S_lycopersicum_Jun_2022.gtf
tomato.config
doIt.sh

Steps for successful run:

1. Add .bash_profile configurations
2. Start a tmux session (tmux new -s base)
3. Start an interactive session (srun --partition Andromeda --cpus-per-task 16 --mem-per-cpu 12000 --time 60:00:00 --pty bash)
4. Activate virtual environment (module load nf-core, module load singularity)
5. Start Nextflow (./doIt.sh SRP252265.csv S_lycopersicum_Jun_2022.fa S_lycopersicum_Jun_2022.gtf S_lycopersicum_Jun_2022.bed tomato.config 1> out.0.txt 2> err.0.txt )
6. Check out file and results directory

Possible error to arise:
Error executing process > 'NFCORE_RNASEQ:RNASEQ:CUSTOM_DUMPSOFTWAREVERSIONS (1)'

Caused by:
Process `NFCORE_RNASEQ:RNASEQ:CUSTOM_DUMPSOFTWAREVERSIONS (1)` terminated with an error exit status (1)

Work dir:
/nobackup/tomato_genome/alt_splicing/SRP252265/work/1e/a6b39f6fb5d57e905cdc43045523cd

Line 52 in versions file:
TRIMGALORE:
trimgalore: perl: error while loading shared libraries: libperl.so: cannot open shared object file: No such file or directory

Discussion: The following error is for a file that contains all of the versions of modules that were used during the pipeline process. The error line 52 in the versions file was for trimgalore. Trimgalore did successfully work with the use of the pipelines resources but on the cluster the module was nonexistent. So the process worked and the results were still there after it finished it was just informing us that the version type couldn't be found for the file to record.

We also created a script to use for easier use (still in the works to be seen as successful):

#!/bin/bash

#SBATCH --time=296:30:00
#SBATCH --nodes=1
#SBATCH --ntasks-per-node=24
#SBATCH --mem=401gb
#SBATCH --job-name=nnnnfcore-splice
#SBATCH --partition=Draco
#SBATCH --output=rrr-%x.%j.out
#SBATCH --error=rrr-%x.%j.err
#SBATCH --mail-type=END,FAIL
#SBATCH --mail-user=mdavi258@uncc.edu

umask 007
set -eu

#file=$(sed -n -e "${SLURM_ARRAY_TASK_ID}p" /nobackup/tomato_genome/rnaseq-phase1/trinity/names.txt)
#dir=$(sed -n -e "${PBS_ARRAYID}p" /lustre/groups/lorainelab/data/illumina/sweet_potato/filtered/dir.txt)

echo "Launching Nextflow NF-Core"

#module load java8 
module load singularity
module load nf-core

##Always use slurm
export NXF_EXECUTOR=slurm
##Save Singularity containers in home dir
export NXF_SINGULARITY_CAHCEDIR="$HOME/nxf"
## Tell nextflow not to use internet
export NXF_OFFLINE='TRUE'
## Control Java Heap Size
export NXF_OPTS="-Xms2g -Xmx8g"

###Change to correct directory with fastq and symbolic link files
cd /nobackup/tomato_genome/alt_splicing/SRP328042-molly

#makeblastdb -in  /projects/tomato_genome/db/reference-4.0/ITAG4.0_cDNA.fasta -dbtype nucl

#nextflow run nf-core/rnaseq –profile ../../test,singularity

#/nobackup/tomato_genome/nfcore_rnaseq/tomato.config



###Change to correct names in directory
nextflow run nf-core-rnaseq-3.4/workflow \
         -resume \
         -profile singularity \
         -c tomato.config \
         --aligner star_salmon \
         --save_trimmed \
         --fasta S_lycopersicum_Jun_2022.fa \
         --input SRP328042.csv \
         --gtf S_lycopersicum_Jun_2022.gtf  \
         --gene_bed S_lycopersicum_Jun_2022.bed \
         --skip_biotype_qc \
         --skip_markduplicates \
         --skip_bigwig \
         --skip_stringtie \
         --skip_qualimap \
         --skip_fastqc


echo "Nextflow Pipeline Finished"

Troubleshooting:

• Need to figure out if it would work better in Andromeda than Draco partition.
• Using the script seems so have more random errors that do not make sense than using the srun which feels like a more organized and precise system to use.

Nextflow nf-core/rnaseq pipeline update:

SRP328042 successfully ran through the nf-core pipeline.
Working Directory:
/nobackup/tomato_genome/alt_splicing/SRP328042-molly

Troubleshooting: I skipped custom_dumpsoftwareversions in the doIt.sh script:
nextflow run nf-core-rnaseq-3.4/workflow \
-resume \
-profile singularity \
-c $CONFIG \
--aligner star_salmon \
--save_trimmed \
--fasta $GENOMEFASTA \
--input $SAMPLESHEET \
--gtf $GTF \
--gene_bed $GENEBED \
--skip_biotype_qc \
--skip_markduplicates \
--skip_bigwig \
--skip_stringtie \
--skip_qualimap \
--skip_fastqc \
--skip_custom_dumpsoftwareversions

During our meeting today, we looked over the files. Everything looks great, so moving this ticket to Done.

Summary:

Data processing output files are on the HPC cluster here:

• /nobackup/tomato_genome/alt_splicing/SRP328042-molly
• /nobackup/tomato_genome/alt_splicing/SRP252265

Planning to align data from two new possible possible controls: SRP100604 and SRP268884