Exercise 4 : Gene prediction

Objective: identify genes in genomic sequences

We will use three different approaches to identify genes within a chicken genomic sequence . Compare the results of each method.

First of all, it is necessary to identify and mask repeated sequences:

• RepeatMasker (return format: html) ( ***    *** )
Mirror sites: Germany, USA, UK (Zebrafish specific)

1- Methods ab initio

• Grail ( ***  *** )
• Genscan ( *** *** )Mirror sites: Genscan (Heidelberg) , Genscan (Paris)

2- Search for transcribed regions (EST, cDNA) in genomic DNA

• Search for ESTs matching the genomic sequence (masked): megablast (NCBI)  (use database est_others) ( *** )
• Select all matching ESTs (at least 95-98% identity, > 70 bp) from chicken (Gallus gallus), and retrieve sequences from NCBI BLAST output  ( *** )
• Assemble ESTs with CAP3  ( ***    *** ) (save Contigs in a text file)
• Align each cDNA to the genomic DNA with  SIM4 ( ***    ***  *** )

 

3- Comparative approach: search protein coding regions by similarity

• Search for proteins matching the genomic sequence (masked) with BLASTX (translated BLAST searches): BLAST (NCBI)( *** )
• Retrieve the closest homologue ( *** )
• Alignment this protein to the genomic DNA with  GENEWISE (Sanger), or GENEWISE (Pasteur) (parameter: genewise protein to genomic DNA)( *** )

A few slides on gene prediction

---

Now repeat the same exercise with two other chicken genomic sequences: genomic1   genomic2  (see results )