It's not clear how to interpret the qualimap (2.2.2b) report. The total number of mapped reads reported by qualimap seems to be <~ 10 % of the mapped reads, and the number of reads reported by qualimap as mapping to "Intergenic" features are < ~10% of the number of reads that featureCounts reports mapping to "Unassigned_NoFeatures", even though these would be expected to be similar values.
It's not clear how to interpret the qualimap (2.2.2b) report. The total number of mapped reads reported by qualimap seems to be <~ 10 % of the mapped reads, and the number of reads reported by qualimap as mapping to "Intergenic" features are < ~10% of the number of reads that featureCounts reports mapping to "Unassigned_NoFeatures", even though these would be expected to be similar values.
Maybe it's subsampling ?