Mapping onto the Transcriptome, RNA-seq

Question

[asyndeton17]

Hi,

I am new to RNA-seq and am planning on mapping to the transcriptome of Mus Musculus. However, I ran into a few questions. Can either bowtie or HISAT2 be used for mapping onto the transcriptome, or does only bowtie work for this? Also, when creating an index before alignment, instead of inputting the genome, do I just input the cDNA file? Lastly, when doing counts using HTSeq-count, do I input the same GTF file used for that of the genome?

Answering any part of the question would help.

Thanks!

EDIT: As of now, I am only interested in differential gene expression.

Answer 1

[Santosh Anand]

You are mixing-up too many things. RNA-seq can be aligned either by 1) splice aware tools like STAR or 2) by any general genome mapper if you provide the transcript fasta file (what you say cDNA file).

Can either bowtie or HISAT2 be used for mapping onto the transcriptome, or does only bowtie work for this?

From HISAT2 manual (https://ccb.jhu.edu/software/hisat2/index.shtml)

HISAT2 is a fast and sensitive alignment program for mapping next-generation sequencing reads (both DNA and RNA) to a population of human genomes (as well as to a single reference genome).

Next

Also, when creating an index before alignment, instead of inputting the genome, do I just input the cDNA file?

It depends upon whether you are mapping with splice-aware tools or general genome mapper. Splice aware tools will need whole genome (with the gtf file of transcript that works to guide the mapping). General genome mapper will need the transcript (cDNA) fasta.

Lastly, when doing counts using HTSeq-count, do I input the same GTF file used for that of the genome?

The gft -file tells you where the genes/transcripts are. So sure, you are going to use the same gtf as genome.

[asyndeton17]

So would HISAT2 work for aligning to the transcriptome?

The rest makes sense.

[Santosh Anand]

The manual says so:

HISAT2 is a fast and sensitive alignment program for mapping next-generation sequencing reads (both DNA and RNA)

Answer 2

[Rob]

HISAT2 can probably be cajoled into aligning to the transcriptome directly, but it is designed to perform spliced read alignment to the genome directly. Bowtie2, on the other hand, is designed to map reads continuously to the indexed reference. It just happens to be particularly good at dealing with multi-mapping reads, which is why people like using it to align to the transcriptome. However, if you're interested in doing quantification with a known transcriptome in a generally well-annotated organism like mouse, I'd recommend doing quantification of the transcripts directly (e.g. using Salmon). If you then want to do DE, this can be followed with something like tximport to get results into your favorite DE tool (e.g. DESeq2, EdgeR, limma-voom).

Answer 3

[mforde84]

Also you need to be clear on what type of differential expression you are interested in. Namely, gene or transcript level expression. For transcript level, I've had good experiences with kallisto and RSEM.