Featurecounts has more reads than star assigned reads ?

Question

[PK]

Hi all,

I know this question was discussed lot. I went through almost all the questions but im not able find the answer for my question. I have a paired end ribo depleted total RNA seq. Post qc everything was great. Then I mapped at the genome level using STAR.

STAR --runMode alignReads --genomeDir $SCRATCHDIR/star_maping/indexed_files/star_index \
      --readFilesCommand zcat --readFilesIn Read_1.fastq Read_2.fastq\
      --outSAMtype BAM SortedByCoordinate \
      --limitBAMsortRAM 16000000000 \
      --outFilterMultimapNmax 1 \
      --sjdbGTFfile /GRCh38.86.gtf \
      --outFilterMatchNmin 30 \
      --runThreadN 6 \
      --quantMode GeneCounts \
      --outFileNamePrefix  /output/$my_name

After mapping for a BAM

                      Number of input reads |   248315401
                  Average input read length |   126
                                UNIQUE READS:
               Uniquely mapped reads number |   229793775
                    Uniquely mapped reads % |   92.54%

Then I used featureCounts to extract the read count using the following parameter.

featureCounts -F GTF -t exon -g gene_id -s 2 -d 50 -D 800 -B -T 6 -a  GRCh38.86.gtf -o ${my_name}_feature_counts.txt sample.bam

When I had look at the summery files I got really lost.

  Process BAM file /scratch.ssd/praveenkumar/job_19723707.meta-pbs.metac ... ||
||    Paired-end reads are included.                                          ||
||    Assign reads to features...                                             ||
||    Total reads : 459587550                                                 ||
||    Successfully assigned reads : 132295130 (28.8%)                         ||

Summery file.txt

Assigned    132295130 Unassigned_Ambiguity  10281242 Unassigned_MultiMapping    0 Unassigned_NoFeatures 317011178


Unassigned_Unmapped 0
Unassigned_MappingQuality   0
Unassigned_FragmentLength   0
Unassigned_Chimera  0
Unassigned_Secondary    0
Unassigned_Nonjunction  0
Unassigned_Duplicate    0

How come the featureCounts reads are almost twice as star assigned unique reads. Is there any way to find?

Thanks in advance

[Istvan Albert]

when there is a discrepancy in counts, and the counts are close to be double (or half) then the first thing I think of is that the paired-end reads are perhaps reported/counted in single mode

[PK]

The report from star is Uniquely mapped reads number | 229793775. How star reports for the paired end library. Featurecounts total reads are exactly twice as 229793775*2 = 459587550. can you please verify my featurecounts parameter. Whether i made any mistake or not. For example featurecount reported me paired-end NO but in the next line it says

Paired-end reads are included

[Pierre Lindenbaum]

please review all the questions you asked during the last seven years and validate them if someone gave the correct answer. Thanks.

Answer 1

[GenoMax]

Featurecounts total reads are exactly twice as 229793775*2 = 459587550.

Since you have a paired-end library you should include the following two parameters in your featureCounts command line.

  -p                  If specified, libraries are assumed to contain paired-end
                      reads. For any library that contains paired-end reads, the
                      'countReadPairs' parameter controls if read pairs or reads
                      should be counted.

  --countReadPairs    If specified, fragments (or templates) will be counted
                      instead of reads. This option is only applicable for
                      paired-end reads. For single-end data, it is ignored.

[PK]

Thanks for your suggestion, i quickly ran the analysis. Yes, you are correct when i included the -p

   Total fragments : 229793775                                                ||
||    Successfully assigned fragments : 95144449 (41.4%)                      ||