基因组注释和组装

Next-generation sequencing technologies produce fragmented data from thousands of short DNA fragments ranging from 50-1000 base pairs, called reads. These reads must be assembled to reconstruct the complete genome sequence in a process called genome assembly. This reassembly is essential because no current sequencing method can sequence an entire genome in one continuous sequence.

Genome assembly involves four main steps: raw data analysis to assess quality and remove contamination or poor-quality reads; contig assembly using overlapping DNA segments to create contiguous sequences; scaffolding using paired reads to order contigs; and gap closing to fill spaces between contigs with sequences or unknown nucleotides.

Comparative genome assembly uses a reference genome from a closely related organism to direct reconstruction, aligning reads to this reference for layout guidance. De novo assembly, performed without a reference genome, relies on overlapping reads to orient sequences into longer contigs. De novo assembly is more challenging but necessary when no suitable reference exists.

Genome annotation identifies functional elements on the assembled genome through two main aims: structural annotation identifies genomic elements like coding regions and regulatory motifs, while functional annotation determines the biological function of these elements, especially protein-coding genes. Annotation tools use known transcripts, protein sequences, and conserved domain signatures as references.

Bacterial genes are not always located in the same genomic positions, and multiple copies of the same gene may appear in different locations. This variability adds complexity to bacterial genome assembly compared to human genomes, where gene physical locations remain constant despite variable copy numbers and repeated sequences within populations.

Long reads exceeding 1 kilobase in length can stitch contigs together more effectively than short paired reads. While short reads leave gaps filled with unknown nucleotides (Ns), long reads allow gaps to be filled with actual DNA sequences, producing more complete and accurate genome assemblies with fewer ambiguities.

Published genomes are updated as sequencing technologies advance and assembly and annotation tools improve. For example, the human genome build 37 released in 2009 was updated to build 38 in 2013. Additionally, genome annotation tools have evolved to provide higher resolution, from annotating long protein-coding genes to individual nucleotides within populations.