alignment software that does not require a reference genome

Question

[regulatinggeneexpression]

Hello, I am trying to do a RNA-Seq on a mouse cell line derived from a mouse strain that is different from the strain used to generate the reference genome (mm9 or mm10).

Is there an alignment software that does need a reference genome? I checked Kallisto but it does not generate BAM files. Thanks.

[Juke34]

"that need a reference genome" Or "does not need" ?

Genome guided assembler : Cufflinks, Stringtie, bayesembler, etc... => sam/bam output

De-novo assembler : Trinity, Oases, SOAPdenovo-Trans, etc... => fasta output

[Kevin Blighe]

This sounds like a good job for HISAT2 / StringTie, but by using the pre-existing reference genome FASTA and transcriptome GTF as guides. This would allow you to see the intricate differences in transcript expression between your strain and the reference.

[Kevin Blighe]

I checked Kallisto but it does not generate BAM files

Kallisto does produce BAMs, or, should I say 'pseudo-bams'.

Answer 1

[h.mon]

You will need to clarify what you want. I understood you want to align using a reference transcriptome and obtain a bam mapping, though it is not clear if a reference transcriptome built from your own reads, or one derived from the standard reference genome and annotation. If you do not want to align to a transcriptome and instead want something else, you will have to spell out clearly what you want.

Do you want to align to a reference transcriptome, and obtain bam files? Salmon can do that, as BWA or Bowtie.

Do you want to assemble a reference transcriptome (with Trinity / Velvet+Oases / etc) from your reads, align the reads to it and obtain a bam file? Again, Salmon, BWA or Bowtie can do that.

With STAR, you can align to a reference genome and output the mapping to the transcriptome.