I am working with tagseq data that I am aligning to a reference transcriptome of an organism. I used several approaches to produce a transcriptome: using Braker annotation pipeline (with reference genome and RNAseq data), Maker annotation pipeline (also with reference genome and RNAseq data) and de novo SPAdes RNA assembler (with RNAseq data only). Braker showed much higher Busco scores for annotated transcriptome than SPAdes assembly (so did Maker). However, when I align tagseq reads to the Braker or Maker transcriptome I get less than 20% of reads alignment rate. When I align the same reads to SPAdes assembly, the alignment rate is ~80%. This tells me something is wrong with the annotation (regardless of the tool - Braker or Maker). Has anyone run into a similar problem or has ideas why I am getting such a different results for predicted vs assembled transcriptomes?
Annotated transcriptomes are based on reference genome and RNAseq data, which may lead to a more refined set of transcripts. Reference bias can result in misannotated or missing transcripts, while SPAdes assembly does not rely on a reference genome, potentially capturing a broader range of transcripts. Alternative transcripts may be more comprehensively captured by SPAdes assembly. Intron handling can lead to more fragmented transcripts, while transcription noise may filter out low-expression or noisy transcripts, resulting in a cleaner but less comprehensive transcriptome.
- Badges
- Science topic
- Similar topics
- Psychology
- Mental Processes
- Learning