Quantifying in mapping-based mode. There is an alternative alignment-based mode where one can align raw reads with another mapper and supplies the alignment (BAM format) in transcript coordinates to Salmon, you can read more about this here
We are using --libType ISR as the sequencing libraries of this dataset were prepared following the Illumina TruSeq Stranded Total RNA protocol. Read here about choosing the appropriate libType
salmon quant is run with 6 threads --threads 6
$cd/home/USER/SSAPs$ declare -a runname=("ERR2675454" "ERR2675455" "ERR2675458" "ERR2675459" "ERR2675460" "ERR2675461" "ERR2675464" "ERR2675465" "ERR2675468" "ERR2675469" "ERR2675472" "ERR2675473" "ERR2675476" "ERR2675477" "ERR2675478" "ERR2675479" "ERR2675480" "ERR2675481" "ERR2675484" "ERR2675485")
for id in ${runname[@]}; do trim1=trimmed/${id}_1.fastq.gz trim2=trimmed/${id}_2.fastq.gzsalmonquant--threads6 \--index/home/USER/db/refanno/gencode.v33_decoys_salmon-1.2.1 \--libTypeISR \--gcBias \--outputsalmon/$id \--mates1 $trim1 --mates2 $trim2done
Inspired by kallisto, Salmon also provides the ability to compute bootstrapped abundance estimates. Such estimates can be useful for downstream (e.g. differential expression analysis) tools that can make use of such uncertainty estimates (e.g. sleuth).
Bootstrap can be enabled by passing the --numBootstraps N option and a positive integer that dictates the number of bootstrap samples to compute. The more samples computed, the better the estimates of varaiance, but the more computation (and time) required.
$cd/home/USER/SSAPs$ declare -a runname=("ERR2675454" "ERR2675455" "ERR2675458" "ERR2675459" "ERR2675460" "ERR2675461" "ERR2675464" "ERR2675465" "ERR2675468" "ERR2675469" "ERR2675472" "ERR2675473" "ERR2675476" "ERR2675477" "ERR2675478" "ERR2675479" "ERR2675480" "ERR2675481" "ERR2675484" "ERR2675485")
for id in ${runname[@]}; do trim1=trimmed/${id}_1.fastq.gz trim2=trimmed/${id}_2.fastq.gzsalmonquant--threads6 \--index/home/USER/db/refanno/gencode.v33_decoys_salmon-1.2.1 \--libTypeISR \--gcBias \--numBootstraps100 \--outputsalmon-bs/$id \--mates1 $trim1 --mates2 $trim2done
kallisto quant is run with 6 threads -t 6 and --rf-stranded as the appropriate library type
Bootstrap is enabled by passing the -b N option and a positive integer that dictates the number of bootstrap samples to compute
Unlike Salmon, Kallisto does not create the top-level folder containing sample-specific outfiles, hence we need to create the top-level folder kallisto before running kallisto quant
$cd/home/USER/SSAPs$ declare -a runname=("ERR2675454" "ERR2675455" "ERR2675458" "ERR2675459" "ERR2675460" "ERR2675461" "ERR2675464" "ERR2675465" "ERR2675468" "ERR2675469" "ERR2675472" "ERR2675473" "ERR2675476" "ERR2675477" "ERR2675478" "ERR2675479" "ERR2675480" "ERR2675481" "ERR2675484" "ERR2675485")
mkdirkallistofor id in ${runname[@]}; do trim1=trimmed/${id}_1.fastq.gz trim2=trimmed/${id}_2.fastq.gzkallistoquant-t6 \-i/home/USER/db/refanno/gencode.v33_kallisto-0.46.2 \--rf-stranded-b100 \-okallisto/$id $trim1 $trim2done