Re: Question about MAKERI usually do mapping with TopHat and Cufflinks. There are two scripts in MAKER to change their output into GFF3 format (tophat2gff and cufflinks2gff). You can then concatenate the GFF3 files and give the file to MAKER in the est_gff field of the maker_opt.ctl file. You can use the resulting mRNA-seq aware annotations to further train SNAP and Augustus to get even better results.
After running MAKER, try viewing the resulting GFF3 output files in Apollo. Sometimes Tophat calls spurious splice site junctions, so after an initial run, you might consider filtering TopHat juctions to only include those with a certain coverage threshold. Also be aware that cufflinks tends to merge exons separated by short introns, so if this is occurring with the genome you are annotating, you can get improved results a by also assembling the mRNAseq data with something like the the celera assembler ( http://sourceforge.net/apps/mediawiki/wgs-assembler/index.php?title=Main_Page) and including the resulting fasta file in the run under the est option of the maker_opt.ctl file (i.e. have both GFF3 and FASTA EST inputs to the same run).
On 9/6/10 1:24 PM, "Tae-Hyuk Ahn" <ahn@...> wrote:
I am a graduate student at Virginia Tech. I have a question about MAKER.
We have transcriptome short sequences by Illumina. We want to use MAKER
to predict/annotate new genes with these sequences.
What is a general way to use mRNA sequences by MAKER?
1> Mapping (e.g., Bowtie)
2> Generating possible exon/intron to GFF (e.g., tophat)
3> Put the result to "EST GFF " of "MAKER"
Do you have any ideas/comments for using mRNA short sequences to find
Thank you in advance!
Department of Computer Science
Virginia Polytechnic Institute & State University
Blacksburg, VA 24061-0106 ahn@...