How does featurecounts deal with the output from Bowtie?

[wang...@gmail.com]

Hi there,

If the following command line is used to generate the SAM file.

bowtie -q -v 0 -k 10 -S -t <index.name> <input.fastq> <out.sam> 

featureCounts program will be used to count the reads by enabling the multi-mapping reads counting:

featureCounts -t miRNA -g Name -O -M -a <mirbasefile> -o <outfile> <samfiles> 

I wonder how featureCounts with the tags such as -O and -M will hand this situation as we know that this SAM file doesn't have "NH" tag? Will the command line above work properly for the microRNA quantification?

Many thanks,

Tom

[Yang LIAO]

Hi Tom,

Because you used the "-M" option for featureCounts, it will try counting all reads no matter if they are multi-mapping or not. It won't check the "NH" tag if multi-mapping reads are allowed. 

MircoRNA reads are usually much shorter, hence there is a higher chance of multi-mapping, and the "-M" option is appropriate. 

Cheers,

Yang

[wang...@gmail.com]

Hi Yang,

Thanks for the answer.

I have some further questions regarding this.

It has been reported that one miRNA might have multiple locations in the genome.

For example, hsa-miR-6859-5p have four locations in the human genome as follows.

1) chr1: 17409 - 17431

2) chr1: 187931 - 187953

3) chr15: 101973529 - 101973551

4) chr16: 17092 - 17114

The questions:

1) When miRNA is considered as meta-feature, will -O be recommended for featureCounts? Why?

2) Should the flag −−fraction be used for featureCounts for this case?

3) If one read from hsa-miR-6859-5p has multiple hits on the above-mentioned four locations that are all defined as the hsa-miR-6859-5p meta-feature, how many hits will be counted for hsa-miR-6859-5p?

4) The Rsubread manual reads "Note that if a read hits a meta-feature, it is always counted once no matter how many features in the meta-feature this read overalps with." Does this apply to the multi-mapping reads? Is one multi-mapping read counted as one read or many different reads?

5) I realise that Subread aligner has the function to map microRNA sequencing reads against the genome. Will there be a massive difference in doing so between Subread aligner and Bowtie? Which one would you suggest?

Many thanks,

Tom