Hi everyone!
I want to map raw reads to a transcriptome and obtain not TPMs, but raw counts. Kallisto, Sailfish and Salmon give an output in TPMs. Featurecounts is not recommended to use with reference transcriptome. It's also impossible to convert TPMs to raw counts.
Is there still a way to obtain them?
Salmon, kallisto and Sailfish all output transcript read counts, more specifically, estimated read counts, as these tools use an expectation-maximization algorithm to estimate optimal counts allocation for multi-mapping reads. The manuals from these three tools clearly state they output read counts. For example, the Salmon manual has this explanation about the NumReads column from the quant.sf file:
NumReads — This is salmon’s estimate of the number of reads mapping to each transcript that was quantified. It is an “estimate” insofar as it is the expected number of reads that have originated from each transcript given the structure of the uniquely mapping and multi-mapping reads and the relative abundance estimates for each transcript.
It is not impossible to convert TPM to counts, if one have the total number of assigned (counted) reads.
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
Bowtie2 or BWA will align raw reads to a reference and output number of physical reads aligned (not pseudo-counts).
The counts are not for each transcipt though. I suppose you are referring to total number of reads aligned, which is not what OP wants.
Yes they are. Why wouldn't they be? If your reference file is separated by contig/transcript, then you can obtain reads mapped to each of these contigs/transcripts
Excuse me, are you sure that Bowtie2 and BWA can map to the reference transcriptome?... I work in Galaxy and there only mapping to the reference genome is suggested.
Aligners can map to any reference provided. If that is logical or not needs to be decided by end-user.
One more question: I have the aligned with the Bowtie2 or BWA reads, and summary for the alignment. How can I make counts out of it? I read that featurecounts is not good for counting reads, aligned to a reference transcriptome.
can you ask your question a different way? I don't understand. Bowtie2 will give you a .sam file, but you need to convert to .bam using samtools to get a counts table.
Extract read counts from fastq files
If you mapped the reads against a reference transcriptome, you can then count reads mapped to each transcript with RSEM, eXpress or Salmon. RSEM is very strict with its alignment requirements, and I believe it won't work with BWA bam files.