1. Predict Genes ab initio

Ab initio prediction means that no other input is used than the target genome itself. Predict the genes in the range 7,000,001-7,500,000 of chr2R of D. melanogaster. Use the FASTA format file chr2R.fa, which includes the whole chromosome 2R.

augustus --species=fly --predictionStart=7000001 --predictionEnd=7500000 chr2R.fa > augustus.abinitio.gff   # takes ~1m

If you want the protein sequences you can retrieve them with

getAnnoFasta.pl augustus.abinitio.gff

2. Predic Genes Using Hints

augustus --species=fly --predictionStart=7000001 --predictionEnd=7500000 chr2R.fa \
  --extrinsicCfgFile=extrinsic.bug.cfg --hintsfile=hints.gff > augustus.hints.gff

3. Creating hints from ESTs or assembled RNAseq transcripts with BLAT

4. Creating hints from proteins with Exonerate

5. Run AUGUSTUS predictions parallel

6. RNAseq integration (raw reads)

6.1 Incorporating RNAseq data into AUGUSTUS predictions with BLAT (including iterative mapping)

6.2 Incorporating RNAseq data with GSNAP (including iterative mapping)

6.3 Incorporating RNAseq data with Bowtie/Tophat (including iterative mapping)

6.4 Incorporating RNAseq data with Palmapper?

7. Training AUGUSTUS