High-throughput per-base count

Question

[Bing He]

What is the exact definition for per-base count in high-throughput sequencing like ChIP-seq, DNase-seq? If one read is of length 20, this read will be counted 20 times for each base covered? I am new to Bioinformatics and Biostars. I searched on the platform, but didn't find answers. Could you please help? Thanks!

I came across this term "per-base count" from this paper, which indicates there is a per-base count c_i for each base i.

Answer 1

[Istvan Albert]

Indeed, isn't it strange that one can publish a paper on how to correct per base counts without ever defining what per base count actually is.

In CHIP-seq some people only use 5' ends of reads but for RNA-Seq everyone uses the full read so your guess of the definition of per-base-readcount is correct.

[Bing He]

Thanks, Istvan! What do you mean by using 5'ends of reads? Do you mean those reads that are mapped to the positive strand?

[Istvan Albert]

It means the rightmost base of the template, 5' end of a read on the positive strand and 3' end of a read on the reverse strand.

Basically to try to reconstruct the borders of the original DNA template.

[Bing He]

Got it. Thanks!