Running kallisto
If a transcriptome sequence set is available for the organism, kallisto can be used for transcript quantification. The kallisto version I used was 0.43.1, which was previously installed using Homebrew by the command like brew install kallisto.
kallisto index -i index transcriptome.fa
kallisto quant -i index -o results/ -t 4 test_1.fq test_2.fq
This command above is for paired end sequence.
The command can accept gzipped FASTA file for indexing.
kallisto index -i index transcriptome.fa.gz
I first tried bzipped(bzip2) FASTQ files for quant, but it failed. FASTQ files must be uncompressed.
For single end, fragment length must be specified by -l option for runnging kallisto quant.
kallisto quant -i index -o results/ -t 4 --single -l 100 -s 20 test.fq
IDs for genes (based on IDs in transcriptome.fa) might be different from those by alignment-based RNA-seq (from GFF file). ID conversion should be done locally before the comparison of those.
An example for batch script to run kalllisto for gzipped-FASTQ files in current directory(single-end read) with bootstrap 100 times (-b 100).
1 2 3 4 5 6 7 8 9 10 | #!/bin/sh
p=4
ref=kallisto_index
outdir=kallisto_outdir
# should be used in the fq containing directory!
for fq in *.fq.gz;
do g="${fq%.fq.gz}"
# for single-end reads. '-l' and '-s' should be modified to fit your data
time kallisto quant -i $ref -o $outdir/$g -t $p --single -l 100 -s 20 $fq -b 100
done
|