Extremely low assigned reads in featureCounts

Question

[xqyjxau]

Hi!Talents! Here are my question:I used my RNA-Seq data(fastq format of human sample). I chose the GRCh38 and GRCh38.89.gtf for STAR reference genome indexing. And later I use featureCounts to count the reads(same gtf file of GRCh38.89.gtf). Here is my command:

featureCounts -T 5 -a <gtf file location> -t exon -g gene_id -o tableCounts *.sam.

And I seem to get extremely low assigned reads: from 18% to 20%.Where is something wrong there?

//================================= Running ==================================\\
||                                                                            ||
|| Load annotation file 
||    Features : 1193949                                                      ||
||    Meta-features : 58233                                                   ||
||    Chromosomes/contigs : 47                                                ||
||                                                                            ||
|| Process SAM file Sample1.              ||
||    Single-end reads are included.                                          ||
||    Assign reads to features...                                             ||
||    Total reads : 20566972                                                  ||
||    Successfully assigned reads : 4086564 (19.9%)                           ||
||    Running time : 0.10 minutes                                             ||
||                                                                            ||
|| Process SAM file Sample2_12AH_0022_Aligned.out.sam...                      ||
||    Single-end reads are included.                                          ||
||    Assign reads to features...                                             ||
||    Total reads : 22822591                                                  ||
||    Successfully assigned reads : 4176597 (18.3%)                           ||
||    Running time : 0.11 minutes                                             ||
||                                                                            ||
|| Process SAM file Sample4_11NH_0240_tumor_Aligned.out.sam...                ||
||    Single-end reads are included.                                          ||
||    Assign reads to features...                                             ||
||    Total reads : 19564290                                                  ||
||    Successfully assigned reads : 2993070 (15.3%)                           ||
||    Running time : 0.10 minutes                                             ||
||                                                                            ||
||                         Read assignment finished.                          ||
||                                                                            ||
|| Summary of counting results can be found in file "tableCounts"             ||
||                                                                            ||
\\===================== http://subread.sourceforge.net/ ======================//

[Istvan Albert]

View your file in IGV and see where your reads align to. If many reads align outside of coding regions then that will be a problem.

Do also a flagstat on your alignment file to see how many reads align overall. That again can be an explanation for the low count.

[Arindam Ghosh]

if "many reads align outside of coding regions" should I proceed with what ever counts I get or should I discard the sample?

[h.mon]

If you are using STAR to map, you can get the counts directly from it by using --quantMode GeneCounts.

[Niek De Klein]

Was the sequencing done strand-specific?

[xqyjxau]

No.It's not strand-specific.