Hi,
I have a question regarding my PE stranded RNA-seq in mouse.
I use HISAT2 for alignment keeping it at “unstranded” (I also tried reverse and forward) and tried to use featurecounts (FC).
Knowing that infer experiment gives me :
This is PairEnd Data
Fraction of reads failed to determine: 0.0132
Fraction of reads explained by “1++,1–,2±,2-+”: 0.0104
Fraction of reads explained by “1±,1-+,2++,2–”: 0.9764
Then both reverse and forward FC give me 0 assigned reads (I kept everything at default)
while the unstranded gives me almost everything assigned.
|Assigned|0|
|Unassigned_Unmapped|954980|
|Unassigned_Read_Type|68588526|
Vs
|Assigned|54328649|
|Unassigned_Unmapped|954980|
|Unassigned_Read_Type|0|
|Unassigned_Singleton|0|
|Unassigned_MappingQuality|0|
|Unassigned_Chimera|0|
|Unassigned_FragmentLength|0|
|Unassigned_Duplicate|0|
|Unassigned_MultiMapping|9082152|
|Unassigned_Secondary|0|
|Unassigned_NonSplit|0|
|Unassigned_NoFeatures|4600484|
|Unassigned_Overlapping_Length|0|
|Unassigned_Ambiguity|577241|
Now interestingly I tried to use the option “count fragments instead of reads” enabled and now the FC reverse only gives me something similar (about twice less, even more than twice less).
Should I keep it like this? I actually get 27 million reads which seems alright because I am supposed to have about 30 million PE reads.
But then I wonder why default for PE is “disabled”.
thanks for your help on this, I’d rather not continue analysis unless I’m sure that part is done right.
best regards