Multiple Mappings

[polsum]

Hi - I have shortRNA NGS data and there are many short reads that map to multiple types of microRNAs and some microRNAs and genes. To which type of microRNA or gene should I assign this read?

For example if a read A maps to 4 different microRNAs mir-1, mir-2 mir-3 and mir-4, should I consider that read as mir-1 or 2 or 3 or 4? or should I equally distribute the mapping? like 25% to each?

Please advice - thanks in advance.

[polsum]

No one? I am pretty sure some of you might have encountered same question or problem. Thanks in advance.

[cjp]

I'm not exactly sure what you mean - but in this cufflinks help page:

http://cufflinks.cbcb.umd.edu/howitworks.html#hmul

they say this for estimating FPKM values when a read maps to multiple places:

"Individual reads will sometimes be mapped to multiple positions in the genome due to sequence repeats and homology. By default, Cufflinks will uniformly divide each multi-mapped read to all of the positions it maps to. In other words, a read mapping to 10 positions will count as 10% of a read at each position."

Chris

[gringer]

Individual reads will sometimes be mapped to multiple positions in the genome due to sequence repeats and homology. By default, Cufflinks will uniformly divide each multi-mapped read to all of the positions it maps to. In other words, a read mapping to 10 positions will count as 10% of a read at each position.

And if you read a little bit further:

If multi-mapped read correction is enabled...Cufflinks will first calculate initial abundance estimates for all transcripts using the uniform dividing scheme. Cufflinks will then re-estimate the abundances dividing each multi-mapped read probabalistically based on the initial abundance estimation of the genes it maps to, the inferred fragment length, and fragment bias (if bias correction is enabled).

[polsum]

Thank you very much cjp, this is exactly what I am looking for.