Alternative mRNA transcripts in Zebrafish

Aim

To discover novel gene models and peptides from alternative splicing that haven’t been annotated yet

The idea

• combine genomics and proteomics
• explore for novel gene models
• identify novel proteins
• search for novel peptide evidence in novel transcripts

Basic Steps

• RNA-Seq experiments
• map RNA-Seq reads against genome and transcriptome to identify junctions (RUM)
• translate mRNAs to proteins to create a novel protein database from RNA-Seq
• mass spectometry experiments
• identify novel proteins (mascot)
• identify and justify novel peptides (custom pipeline)

Data

• post fertilized developing zebrafish embryos

• for both RNA-Seq and MS2

  • 1 day experiment with 4 replicates
  • 3 days experiment with 3 replicates
  • 5 days experiment with 5 replicates

• mrnas from UCSC (lineformat)

• proteins sequences from UCSC (fasta)

• genome sequences (fasta)

• RNA-Seq short reads

• MS2 data that contain peptide spectrum matches (mgf)

• high quality junctions from RUM (bed)

• protein/peptide matches from mascot

Steps

• create protein database for mascot searches
• create novel proteins
• identify peptides of the novel proteins
• double check by searching for novel peptides in a database that contains both annotated and novel proteins
• visualize the candidates in UCSC genome browser

Pipeline

• clean the raw db files (only before creating the db on mascot server) ruby scripts/clean_files.rb data/mrnas_ucsc_lineformat.txt data/proteins_ucsc_db.fa data/danRer7.fa results/mrnas_ucsc_lineformat_clean.txt results/proteins_ucsc_db_clean.fa

• find novel proteins ruby scripts/find_unannotated_proteins.rb results/mrnas_ucsc_lineformat_clean.txt data/1d-30oct-zfish-ucsc-db.dat results/proteins_ucsc_db_clean.fa data/junctions_high-quality.bed data/danRer7.fa results/1d-30oct/peptides-1d-30oct.bed results/1d-30oct/mrnas_unannotated.bed results/1d-30oct/mrnas_unannotated_seqs.fa results/1d-30oct/mrnas_unannotated_lineformat.txt results/1d-30oct/proteins_unannotated.fa

• run mascot for novel proteins database

• create bed file with the peptides of the novel proteins ruby scripts/peptides_to_bed.rb results/1d-30oct/mrnas_unannotated_lineformat.txt data/1d-30oct-zfish-unannotated-db.dat results/1d-30oct/proteins_unannotated.fa results/1d-30oct/peptides_on_unannotated_proteins-1d-30oct.bed

• create the db on mascot server

• join the db files (mrnas_lineformat and proteins) cat results/mrnas_ucsc_lineformat_clean.txt > results/mrnas_full_db_lineformat.txt grep -v "#" results/mrnas_unannotated_lineformat.txt >> results/mrnas_full_db_lineformat.txt cat results/proteins_ucsc_db_clean.fa results/proteins_unannotated.fa >> results/proteins_full_db.fa

• run mascot for joined db

• find novel peptides with evidence ruby scripts/find_unexplained_peptides.rb data/1d-30oct-zfish-full-db.dat results/mrnas_full_db_lineformat.txt data/mrnas_ucsc_lineformat.txt results/1d-30oct/unexplained_peptides_of_unannotated_proteins-1d-30oct.bed results/1d-30oct/unique_peptides_of_unannotated_proteins-1d-30oct.bed results/1d-30oct/nonunique_peptides_of_unannotated_proteins-1d-30oct.bed results/1d-30oct/unexplained_peptides_of_unannotated_proteins_per_spectra-1d-30oct.bed

TODO list

• review the peptide candidates
• add more unit tests
• run all datasets
• TPP for data validation
• visualization of genomic annotation information with DAS