Bowtie2 SAM output missing base quality scores

[Comosus]

I've been using the latest version of Bowtie2 to map 100 base Illumina reads for RNA-seq. I am also trimming 25 bases from the 3' end.

I've mapped this sort of data before with TopHat and haven't come across problems. However, I don't understand certain aspects of the .sam output that Bowtie2 gives me:

1. The output contains ALL reads, whether mapped or not - not too much of a problem since I can filter out the unmapped reads anyway
2. In some cases, the base quality scores are non-existent (just represented by "*"). I've checked the .fastq files that I supplied to Bowtie2, and these do contain the base quality scores for these reads. I thought at first that it was only missing on unmapped reads, but it is also missing on reads that have mapped.

Any help would be greatly appreciated.

[TiborNagy]

I have tried bowtie2 version 2.0.0.beta7 and I did not noticed any bugs with the base quality scores, so maybe you find a new bug. Bowtie2 still in beta, so it is not a big surprise. Try to map the untrimmed reads, just to be sure the error not related to the trimming process.

[Lluc]

I had the same problem, and then realized that the SAM format allows for the string of base qualities to be just "*". I guess it is a feature, to save space. In my case, the only mappings without the base quality string were secondary mappings. When a read was mapped to more than one location, only the primary mapping stored the base qualities; and only secondary mappings were missing the qualities. At least, the information was not lost.

I don't know if there is a way to ask bowtie2 to behave differently, but it would be useful.