I assembled a reference transcriptome de novo using Trinity with 2x300bp Illumina MiSeq PE reads. I used libraries from 4 individuals that were multiplexed over 4 lanes and chose "automated adapter trimming through the CASAVA pipeline." Assuming all of the adapters had been removed, I used a QC pipeline but skipped using 'scythe' to remove 3' end adapter contamination, assuming that these had been automatically trimmed when the samples were demultiplexed.
Since then, I have assembled and annotated the transcriptome (BLAST-style descriptions with usearch, protein prediction with InterProScan, mapped GO terms, etc.). However, as I attempted to upload the Transcriptome Shotgun Assembly (TSA) to NCBI, I received many error messages indicating that my assembly contained adapter contamination. Here is an example of the output:
ERROR:
File: pinus.albicaulis_transcriptome_tsa_v1.fsa, Code(VECTOR_MATCH), Sequence-id: TR2-c2_g1_i1, Interval: 2159..2224, This sequence has a Strong match on the following UniVec vector: gnl|uv|NGB00870.1:1-66 NEBNext Index 27 Primer for Illumina
[My library preparation used TruSeq LT kit which apparently is the same]
After looking into it, it seems that flagged transcripts do indeed have 3' end index adapter contamination and some have AAAAA tails afterward. Of course, I cannot upload the TSA without solving this problem.
Is there any way to simply trim adapter contamination from the completed reference transcriptome .fasta file? Many of the contaminated assembled sequences are even annotated and it seems these adapter sequences are present in their full form. I know adapter trimming programs generally require .fastq quality information, but I'm really hoping there may be an alternative to starting over again from the raw read data.
Is this a massive setback? Will I need to redo everything from scratch? Thanks for any help in advance.
Since then, I have assembled and annotated the transcriptome (BLAST-style descriptions with usearch, protein prediction with InterProScan, mapped GO terms, etc.). However, as I attempted to upload the Transcriptome Shotgun Assembly (TSA) to NCBI, I received many error messages indicating that my assembly contained adapter contamination. Here is an example of the output:
ERROR:
File: pinus.albicaulis_transcriptome_tsa_v1.fsa, Code(VECTOR_MATCH), Sequence-id: TR2-c2_g1_i1, Interval: 2159..2224, This sequence has a Strong match on the following UniVec vector: gnl|uv|NGB00870.1:1-66 NEBNext Index 27 Primer for Illumina
[My library preparation used TruSeq LT kit which apparently is the same]
After looking into it, it seems that flagged transcripts do indeed have 3' end index adapter contamination and some have AAAAA tails afterward. Of course, I cannot upload the TSA without solving this problem.
Is there any way to simply trim adapter contamination from the completed reference transcriptome .fasta file? Many of the contaminated assembled sequences are even annotated and it seems these adapter sequences are present in their full form. I know adapter trimming programs generally require .fastq quality information, but I'm really hoping there may be an alternative to starting over again from the raw read data.
Is this a massive setback? Will I need to redo everything from scratch? Thanks for any help in advance.
Comment