Improving Silkworm Genome Annotation Using a Proteogenomics Approach

The
silkworm genome has been deeply sequenced and assembled, but
accurate genome annotation, which is important for modern biological
research, remains far from complete. To improve silkworm genome annotation,
we carried out a proteogenomics analysis using 9.8 million mass spectra
collected from different tissues and developmental stages of the silkworm.
The results confirmed the translational products of 4307 existing
gene models and identified 1701 novel genome search-specific peptides
(GSSPs). Using these GSSPs, 74 novel gene-coding sequences were identified,
and 121 existing gene models were corrected. We also identified 1182
novel junction peptides based on an exon-skipping database that resulted
in the identification of 973 alternative splicing sites. Furthermore,
we performed RNA-seq analysis to improve silkworm genome annotation
at the transcriptional level. A total of 1704 new transcripts and
1136 new exons were identified, 2581 untranslated regions (UTRs) were
revised, and 1301 alternative splicing (AS) genes were identified.
The transcriptomics results were integrated with the proteomics data
to further complement and verify the new annotations. In addition,
14 incorrect genes and 10 skipped exons were verified using the two
analysis methods. Altogether, we identified 1838 new transcripts and
1593 AS genes and revised 5074 existing genes using proteogenomics
and transcriptome analyses. Data are available via ProteomeXchange
with identifier PXD009672. The large-scale proteogenomics and transcriptome
analyses in this study will greatly improve silkworm genome annotation
and contribute to future studies.