I like ART for NGS. It's comprehensive and one can easily generate custom profiles from a given fastq samples, e-g- MiSeq 300bp or brand new HiSeq 250bp
Default pbsim reads come with a little more errors and shorter reads than what in my experience you get with real-life pacbio data. For convenience I use the script below to directly generate PacBio reads from a reference FASTA with length and error distributions closer to what you get from current P4-C2/P5-C3 runs. (Just put the code in a file, e.g. pbsim-clr, next to the pbsim binary and make it executable)
#!/bin/bash
BIN="$(dirname "$(readlink -f "$0")")";
COV=10;
[[ $1 == '-c' ]] && { COV=$2; shift; shift; };
[[ $# -eq 0 || $1 == "-h"* ]] && { echo -e "Usage:\n pbsim-clr [-c cov] FASTA [FASTA ..]"; exit 0; }
for FILE in $@; do
PRE=`basename ${FILE%.*}-pb`;
$BIN/pbsim \
--data-type CLR \
--depth $COV \
--length-max 50000 \
--length-min 1000 \
--length-mean 7000 \
--difference-ratio 8:62:30 \
--accuracy-mean 0.83 \
--model_qc $BIN/data/model_qc_clr \
$FILE
rm sd_0001.ref
I=1;
OUT=$PRE-`printf %02d $I`
while [[ -e $OUT.fq ]]; do
I=$(( $I + 1 ));
OUT=$PRE-`printf %02d $I`;
done;
mv sd_0001.fastq $OUT.fq
mv sd_0001.maf $OUT.maf
done;
The DAZZLER simulater only produces random sequences with PacBio length profile.
[edit] Edited to prevent any confusion about dazzler - see comments.
