Comparison of assembly algorithms for improving rate of metatranscriptomic functional annotation

Area of Science:

• Microbial genomics
• Bioinformatics
• Metatranscriptomics

Background:

• High-throughput RNA sequencing (metatranscriptomics) enables functional analysis of microbial communities.
• Accurate annotation of short sequence reads to bacterial transcripts is crucial.
• Assembling reads into longer contigs can aid annotation, especially without reference genomes.

Purpose of the Study:

• To evaluate the performance of four assemblers for metatranscriptomic datasets.
• To determine the effectiveness of assembly in improving read annotation.
• To identify the best-performing assembler for complex microbial communities.

Main Methods:

• Evaluation of four assemblers: Trinity, Oases, Metavelvet, and IDBA-MT.
• Utilized single-end and paired-end RNA sequence reads from mouse gut microbiome.
• Assessed performance based on contigs assembled, reads assigned, and transcript annotation rates.

Main Results:

• Trinity assembler demonstrated superior performance in contig assembly and read annotation.
• Trinity assembly increased annotated reads from 15.5% to 50.3%.
• Complex datasets with similar sequences introduced assembly errors, highlighting potential limitations.

Conclusions:

• Metatranscriptome assembly substantially enhances read annotation accuracy.
• Trinity is the top-performing assembler among those tested.
• Future analyses should consider pre-assembly read sorting for complex, homologous datasets.