Featurecounts in built genome giving no read assignments

Hello All,
I have been trying to use featurecounts to count by RNA-seq reads. When I use a downloaded GTF from UCSC - I get the results for transcript IDs not gene IDs, so I tried to use the in-built genome as I did for the HISAT2, however despite everything being set to hg19 - no reads are being assigned and I get 0 counts for all genes. Can anyone advise as to what I am doing wrong please?
Thanks in advance :slight_smile:

Hi @KeeleyB
similar questions were asked on the forum. Search the forum for featureCounts and check suggestions.
Hope that helps.

Hi Igor,
Yes I have been reading your previously answers to post like these - but if I am using the in-built hg19 genomes for alignments how can I check that the alignments are named the same for them to be recognised… I was using in built hg19 in HISAT2 and then in-built one in Feature counts.
Thanks in advance

hi @KeeleyB
what UCSC annotation was used? Maybe try hg19.refGene.gtf.gz. It should produce gene counts with the default featureCounts settings. Usually it is very reliable with single end data.

Do you have paired-end data (trimmed)? Have you used ‘strandness’ setting for HiSAT2 and featureCounts? What do you see in the summary file? Do you see majority of reads in Unassigned_Read_Type category? If yes, featureCounts manual gives the following description:
reads that have an unexpected read type (eg. being a single
end read included in a paired end dataset) and also cannot be counted with confidence
(eg. due to stranded counting). Such reads are typically generated from a read trimming
Make sure you use correct strand settings in HiSAT2 and featureCounts. To determine strandness, map one sample with unstranded settings and run Infer Experiment. Enable Count fragments instead of reads in Options for paired reads (featureCounts).
Hope that helps.
Kind regards,

1 Like