mapping reads against luciferase with bowtie

[mjp]

I have used bowtie to map my reads against Arabidopsis.

I have also built a reference genome out of my luciferase gene sequence and did map my reads against that.

I'm wondering whether building bowtie indexes out of that short sequence will affect my further analysis?

Do you think that after applying the same way of normalization (in my case for total read number and for gene length) I will have comparable results between luciferase and Arabidopsis?

In luciferase case my number of mapped reads will be equal to total reads mapped so doing this normalization does not make a lot of sense.

RPKM maybe?

[colindaven]

Comparing transcripts from one gene against transcripts mapped against a whole genome sounds a little bizarre. Maybe you need to explain what you're trying to achieve in the project ?

[mjp]

Luciferase is the control gene that I have used in my experiment. It is fused with the rd29a promoter. In my experiment luciferase level of expression should be at the very similar level as the rd29a gene as they share the same promoter.

I would like to have a relative level of expression between luciferase and rd29a gene.

I have mapped my reads against Arabidopsis genome so I have rd29a level of expression. Now I'm trying to figure out whether building reference genome with bowtie out of luciferase sequence and performing the mapping of my reads against that would give me something that I could compare with rd29a.

Thanks.

[simonandrews]

Personally I'd have been tempted to put the luciferase into the existing Arabidopsis genome (as an extra 'chromosome') so I could discount any possibility of non-specific matches being different between the two genomes you currently have. It may make no difference, but it will mean that you can get away with only doing a single search if nothing else.

[feixue1039]

Hi,mjp
As I know,only the uniquely mapped reads are considered in the calculation of RPKM value. And a read uniquely map to luciferase gene sequence may map alternatively to many loci of your genome.So I advise you not to build bowtie indexes with the luciferase gene sequence only.
Just my humble opinion.

feixue1039

[mjp]

Thanks guys,

I followed your ideas and built the one set of indices after including my luciferase genes as a separate chromosome. I actually included my entire plasmid as a separate chromosome.

This step did not give me better results but they are more comparable now I think. I also found nice set of reads mapped in different regions of my plasmid which is somewhat backing up the luciferase mappings.

Thanks once again.