Strandedness of RNA-seq results

Question

[jamesymtang]

Hi all,

I am not very familiar with bioinfo things as my background is more towards wetlab.

I am analysing my dual RNA-seq results (host+pathogen RNA), and I am using HISAT2 (aligned to genome) + featurecounts pipeline.

I am not sure why I only get assigned reads from unstrand option, when my cDNA libaries are supposed to be built from (KAPA Stranded mRNA-Seq Kit + xGen® Dual Index UMI Adapters), which should be reverse stranded according to instructions online?

Any help will be appreciated.

Yours Sincerely, James PhD student

[rpolicastro]

What code are you running for HISAT2 and featureCounts?

[jamesymtang]

Hi rpolicastro,

After some more reading, I find out that I should be using F/FR for HISAT2 and -1(Forward) for featurecounts according to (https://rnabio.org/module-09-appendix/0009/12/01/StrandSettings/). However, my problem still remains....

I am not using codes unfortunately but the online platform usegalaxy.

My sequencing structure is paired-end.

In terms of codes, I copied the system message online:

When I do HISAT2 with FR stranded or unstranded mode, I get:

76845532 reads; of these: 76845532 (100.00%) were paired; of these: 3352367 (4.36%) aligned concordantly 0 times 64109899 (83.43%) aligned concordantly exactly 1 time

9383266 (12.21%) aligned concordantly >1 times
----
3352367 pairs aligned concordantly 0 times; of these:
  174177 (5.20%) aligned discordantly 1 time
----
3178190 pairs aligned 0 times concordantly or discordantly; of these:
  6356380 mates make up the pairs; of these:
    3796659 (59.73%) aligned 0 times
    1958907 (30.82%) aligned exactly 1 time
    600814 (9.45%) aligned >1 times

97.53% overall alignment rate

which looks reasonable?

but when I do featurecounts with FR stranded mode:

Status 0U hisat2 combined Assigned 0 Unassigned_Unmapped 4151857 Unassigned_Read_Type 189796595 Unassigned_Singleton 0 Unassigned_MappingQuality 0 Unassigned_Chimera 0 Unassigned_FragmentLength 0 Unassigned_Duplicate 0 Unassigned_MultiMapping 0 Unassigned_Secondary 0 Unassigned_NonSplit 0 Unassigned_NoFeatures 0 Unassigned_Overlapping_Length 0 Unassigned_Ambiguity 0

whereas unstranded I get:

Status 0U hisat2 combined Assigned 108019249 Unassigned_Unmapped 4151857 Unassigned_Read_Type 0 Unassigned_Singleton 0 Unassigned_MappingQuality 0 Unassigned_Chimera 0 Unassigned_FragmentLength 0 Unassigned_Duplicate 0 Unassigned_MultiMapping 53126687 Unassigned_Secondary 0 Unassigned_NonSplit 0 Unassigned_NoFeatures 16477673 Unassigned_Overlapping_Length 0 Unassigned_Ambiguity 12172986

I checked my strandedness with infer experiment according to instructions online:

This is PairEnd Data Fraction of reads failed to determine: 0.4374 Fraction of reads explained by "1++,1--,2+-,2-+": 0.4819 Fraction of reads explained by "1+-,1-+,2++,2--": 0.0808

which should equal to forward stranded?

I have also doublecheck my chromosome annotations, it shouldnt be a problem... so I really don't know why its not working with FR stranded

[Trivas]

Are you sure it's not supposed to be RF? Your read 1 should be the reverse of the RNA if your kit uses the dUTP method.

[Trivas]

Agree with rpolicastro to post your mapping command. What's your sequencing structure (eg paired-end or single-end)?

[charles.feigin]

Hi, coming here for the same reason. Libraries prepared with the Illumina Stranded mRNA Prep kit (dUTP-based).

In hisat2 with --rf I get <5% concordant alignments, whereas in --fr or unstranded I get ~>80% concordant alignments.

When I move to featureCounts though, I get <1% assignment when using -s 1 (which should correspond to --fr in hisat2), while I get >70% assignment when using -s (which should correspond to -rf in hisat2).

There seems to maybe be a mismatch between the parameters of these programs?