RSEM failed on miRNA read dataset

Question

[bangbangphil2]

Hi,

I'm working on a miRNA quantification RNA-Seq pipeline and I'm using RSEM for the quantification step. However some of my read sets are unable to be quantified by RSEM. I don't understands since there are aligned reads by bowtie2 (the aligner used underneath RSEM) but RSEM seems unable to use them for quantification ...

Here is the ouput of RSEM for one problematic read set :

Warning: the seed length set is less than 25! This is only allowed if the references are not added poly(A) tails.
/opt/bioinformatics/bowtie2/2.2.3/bin/bowtie2 -q --phred33 -D 20 -R 4 -N 0 -L 15 -i S,1,0.50 --dpad 0 --gbar 99999999 -p 24 -k 200 -x /net/archive04/mnt/tank/archive/rklinck/phil/rsem_ref/mirbase21_hsa_dna_mature/mirbase21_hsa_dna_mature -U ../fastq-t/E-07.fastq.gz | samtools view -S -b -o E-07.mature.temp/E-07.mature.bam -
[samopen] SAM header is present: 2588 sequences.
4896945 reads; of these:
  4896945 (100.00%) were unpaired; of these:
    4442252 (90.71%) aligned 0 times
    284847 (5.82%) aligned exactly 1 time
    169846 (3.47%) aligned >1 times
9.29% overall alignment rate

rsem-parse-alignments /net/archive04/mnt/tank/archive/rklinck/phil/rsem_ref/mirbase21_hsa_dna_mature/mirbase21_hsa_dna_mature E-07.mature.temp/E-07.mature E-07.mature.stat/E-07.mature b E-07.mature.temp/E-07.mature.bam -t 1 -tag XM
Parsed 1000000 entries
Parsed 2000000 entries
Parsed 3000000 entries
Parsed 4000000 entries
Parsed 5000000 entries
Done!

rsem-build-read-index 32 1 0 E-07.mature.temp/E-07.mature_alignable.fq
Build Index E-07.mature.temp/E-07.mature_alignable.fq is Done!

rsem-run-em /net/archive04/mnt/tank/archive/rklinck/phil/rsem_ref/mirbase21_hsa_dna_mature/mirbase21_hsa_dna_mature 1 E-07.mature E-07.mature.temp/E-07.mature E-07.mature.stat/E-07.mature -p 24 -b b E-07.mature.temp/E-07.mature.bam 0
Refs.loadRefs finished!
Thread 0 : N = 18998, NHit = 27466
Thread 1 : N = 18965, NHit = 27466
Thread 2 : N = 18878, NHit = 27467
[... truncated ...]
Thread 22 : N = 18906, NHit = 27466
DAT 0 reads left
Thread 23 : N = 18804, NHit = 27459
EM_init finished!
1000000 READS PROCESSED
2000000 READS PROCESSED
3000000 READS PROCESSED
4000000 READS PROCESSED
estimateFromReads, N0 finished.
estimateFromReads, N1 finished.
ROUND = 1, SUM = 4896944.99999999, bChange = 850.254, totNum = 2588
ROUND = 2, SUM = 4896945, bChange = 2.21846, totNum = 261
[... truncated ...]
ROUND = 50, SUM = 4896945, bChange = 0.00100884, totNum = 1
ROUND = 51, SUM = 4896945, bChange = 0.000943811, totNum = 0
No alignable reads?!
"rsem-run-em /net/archive04/mnt/tank/archive/rklinck/phil/rsem_ref/mirbase21_hsa_dna_mature/mirbase21_hsa_dna_mature 1 E-07.mature E-07.mature.temp/E-07.mature E-07.mature.stat/E-07.mature -p 24 -b b E-07.mature.temp/E-07.mature.bam 0" failed! Plase check if you provide correct parameters/options for the pipeline!

thanks for your help!

Phil

Answer 1

[Michele Busby]

RSEM isn't really the right tool for this. RSEM is for quantifying transcripts and finding splice junctions and similar tasks. The difficulty with miRNA is the small read size and multiple mapping, so I would look around for a tool that is specialized for that.

See:

https://groups.google.com/forum/#!topic/rsem-users/VA9FWK-QFKs

Answer 2

[igor]

Most RNA-seq tools aren't really designed for very short sequences that you get from small RNA-seq. I would suggest using a miRNA-specific tool, such as:

• miRge - http://atlas.pathology.jhu.edu/baras/miRge.html
• miRDeep2 - https://www.mdc-berlin.de/8551903/en/
• miRExpress - http://mirexpress.mbc.nctu.edu.tw/