Hii Rohan,

Thank you for your reply.

So these files used in Augustus for gene prediction or not.

Dear Maulik, you can use your assembled and unassembled reads to generate hints for Augustus. Then can ab initio train Augustus for gene prediction. Just remember to use the generated hints during final gene model prediction. You will also get the information about the percentage of hints support for you gene models.

Hii Ashmit,

Thank you for your reply,

Please tell me the example commands for the same you explain in your answer

Dear Maulik,

You can refer to the PASA documentation (http://pasa.sourceforge.net/).

Breifly,

1) cleaning the transcript sequences

seqclean transcripts.fasta -v /path/to/your/vectors.fasta

2) Align the transcript sequences to the genome using blat or gmap, and then assembly based on the alignments

../scripts/Launch_PASA_pipeline.pl -c alignAssembly.config -C -R -g genome.fasta -t transcripts.fasta.clean -T -u transcripts.fasta --ALIGNERS blat,gmap --CPU 2

Note:

A template configuration file for alignAssembly.config is provided at: $PASAHOME/pasa_conf/pasa.alignAssembly.Template.txt

Simply copy pasa.alignAssembly.Template.txt to your working directory and set the values for your MySQL database settings.  You only need concern yourself with the following values:

# MySQL settings

MYSQLDB=

#######################################################

# Parameters to specify to specific scripts in pipeline

# create a key = "script_name" + ":" + "parameter"

# assign a value as done above.

#script validate_alignments_in_db.dbi

validate_alignments_in_db.dbi:--MIN_PERCENT_ALIGNED=

validate_alignments_in_db.dbi:--MIN_AVG_PER_ID=

#script subcluster_builder.dbi

subcluster_builder.dbi:-m=50

Hope it will help you.

All the best,

En-Hua

Dear En-Hua,

Thank you for your reply

You give me the nice idea about the creation of Gene model

I have gone through the pipeline it gives the good results.

If you tell me about the Augustus pipeline then it will be better n good for me.

Thank You

Dear Maulik,

Did you mean how to use Augustus to predict the gene models? What you want is just the parameters used or the detail full command?

1) For the parameters, you can run like this:

./augustus --help

The most important parameter as I know is "--species", which will affect your final result greatly. Also you can use the command "augustus --species=help" to choose the best species model for your targer genome.

If there are not any species models closely to your sequenced species, you can train your own models using the RNA-seq data or protein data from the closely species. You can refer to this website (http://www.molecularevolution.org/molevolfiles/exercises/augustus/training.html) to see the detail steps for "species" parameter training.

2) As for the detail command line, I often run like this:

augustus --gff3=on --species=xxx --outfile=your_genome.gff3 your_genome.fasta.masked

That's all I know about Augustus. Hope it will help you.

All the best,

En-Hua Xia

Dear En-Hua,

Thank you very much,

I have gone through the link for training set and it is easy to do with augustus rather than PASA. I have done through the augustus

PASA is very complex and i have not installed yet some error in database side.

In PASA my database connection is not proper.

Thanks again En-Hua.

Dear En-Hua,

I have used PASA as per your steps . I can do properly with good results but today i got the error in PASA DB settings.

Thank You

This conversation looks very much similar to my question. How do i know which species parameter will work the best for my species. I ran augustus with a sample genome first with arabidopsis and next with fly. with fly i got a decent number of models, but with arabidopsis i got 3 fold more models than fly. Can any one explain me how can this happen.