Programs/Parameters For De Novo Transcriptome Assembly Of Ion Torrent Reads?

Question

[Ahdf-Lell-Kocks]

I am evaluating different tools for de novo transcriptome assembly (not reference-based or mapping reads to the contigs approach, I am not interested in such approach). I would like to ask for feedback to see what is people doing that works for them.

I have tried MIRA and SeqMan NGEN so far, but there are many parameters and I am not sure which ones are best. I am working with the example human datasets available at the Ion Torrent Community website (requires authentication), so these are fastq files, not sff/xml files.

For MIRA, I am using this command line:

id=HBR.polyA.WT.DID-143.120711
ln -s $id.fastq ${id}_in.iontor.fastq
~/src/mira/latest/mira/bin/mira \
--project=HBR.polyA.WT.DID-143.120711 -DI:trt=/tmp/ \
--job=denovo,est,accurate,iontor IONTOR_SETTINGS -LR:mxti=no

For SeqMan Ngen, I am using the default parameters:

; Scan Controls
setQualityParam MinAveLowQual: 15
; Reads
loadSeqIonTorrent file: "/my/dir/HBR.polyA.WT.DID-143.120711.fastq"  maxSeqs: 99999999 
; Controls
setParam MatchSize: 15
setParam MatchSpacing: 20
setParam MinMatchPercent: 85
setParam UseRepeatHandling: false
; Actions
assemble trimEnds: true vectScan: false repeatScan: false contamScan: false doAssemble: true

Any suggestions?