Hi,
We did some miRNA sequencing (3 WT vs. 3 condition) and I would like to carry out differential expression analysis using DESeq (DESeq2). The alignment was done using the RNA-Star aligner (part of the ENCODE project). I know that HTSeq-count discards multi-mapped reads, however, for miRNA counts, discarding these reads would not capture the entire picture. A lot of the mature miRNA sequences are identical (especially when the miRNAs belong to the same family) and thus sequencing reads would be mapped to all the corresponding locations. Based on this, if we ignore these reads we are going to be left with zero to very few read counts for differential expression analysis.
HTseq count uses the NH column in the SAM output to look for multi mapped reads and RNA-STAR outputs this column (not all aligners do). So is there a way to turn the multi-mapped filtering OFF in HTSeq or should I use a custom counting script (instead of HTSeq) to generate my count tables. Also, do you think this would be a reasonable way to then use DESeq (we are currently focusing on validating an experimental idea/design and some of the interesting DE miRNAs would then be checked for by using qPCR).
We did some miRNA sequencing (3 WT vs. 3 condition) and I would like to carry out differential expression analysis using DESeq (DESeq2). The alignment was done using the RNA-Star aligner (part of the ENCODE project). I know that HTSeq-count discards multi-mapped reads, however, for miRNA counts, discarding these reads would not capture the entire picture. A lot of the mature miRNA sequences are identical (especially when the miRNAs belong to the same family) and thus sequencing reads would be mapped to all the corresponding locations. Based on this, if we ignore these reads we are going to be left with zero to very few read counts for differential expression analysis.
HTseq count uses the NH column in the SAM output to look for multi mapped reads and RNA-STAR outputs this column (not all aligners do). So is there a way to turn the multi-mapped filtering OFF in HTSeq or should I use a custom counting script (instead of HTSeq) to generate my count tables. Also, do you think this would be a reasonable way to then use DESeq (we are currently focusing on validating an experimental idea/design and some of the interesting DE miRNAs would then be checked for by using qPCR).
Comment