Difference between unique reads and read count

Question

[reswinvented]

I have the results from an microRNA Seq experiment. After inputting the FASTA files into miRanalyzer, I was able to obtain a bunch of plain text files for each of my samples. i.e.

name    #uniqueReads    readCount    norm_expressed_all    norm_expressed_mapped
hsa-miR-122-5p    2529    676685    223871.9557685112    314406.13233320817

What is the difference between unique reads and read count? I am trying to use DESeq2 for DE analysis, and I want to generate a plain text file for each of my sample to feed into DESeq2 as input so I can generate a count table matrix for the analysis, which column should i use? read count or unique reads? Thank you.

Answer 1

[Devon Ryan]

According to the documentation: readCount: the read count sum of all unique reads.

So just use the second column as the count for DESeq2.

Edit: BTW, you may still want to look in the *_ambig file for cases where it makes more sense to look at a miRNA family (if the various members are near-identical, such that you get very few counts, then you end up needing to look at the family as a whole).

[reswinvented]

Thanks! That is very helpful, especially the edit

Answer 2

[David Langenberger]

Unique reads are also called 'tags'. Since the measured mature microRNA sequences arise from a Dicer processing, most of them are cut at exactly the same position and result in exactly the same read sequence (ca. 24nt). Thus, you find these sequences in your fastq file several thousand times.

But now to your question: Take the read count, since the absolute read count is the 'expression' of the microRNA. (as Devon already mentioned)

[reswinvented]

Thanks a bunch!

[wt215]

Thank you for your explanation! That helps a lot!