Cannot understand Tophat output... Help!

[gen2prot]

Hello All,

I built an index using Drosophila genes giving the bowtie-build command. This worked just fine and I got the six files. Then I ran Tophat from the same folder using the following command

tophat --solexa1.3-quals -p 4 GeneIndex path_to_reads

My reads are unpaired. The output says that the Tophat run was OK. There is nothing in the error file. However in the tophat_out folder, I get the following files and directories.

File: GeneIndex.fa
File: left_kept_reads.fq
Folders: log, tmp

Within these folders I do not find the files accepted.sam and junctions.bed. What am I doing wrong. Any suggestions?

Thank you
Abhijit

[gen2prot]

Hello,

Tophat seems to have run properly, the error output has no warnings. However the accepted.sam file shows no hits. The junctions.bed file is 1.3 MB. Anyone faced a similar problem?

Thanks
Abhijit

[gen2prot]

Hello,

Following up on this, I wanted to know if indexing of all genes is possible or not in the first place. My genes.fasta file looks something like this.

>FBgn0034974 type=gene; loc=2R:19969255..19973683; ID=FBgn0034974; name=CG16786; dbxref=FlyBase:FBgn0034974,FlyBase:FBan0016786,FlyBase_Annotation_IDs:CG16786,GB:BI363616,GB:BT001664,GB_protein:AAN71419,GB_protein:AAF47161,GB_protein:AAM68305,UniProt/TrEMBL:Q7JRF0,INTERPRO:IPR011071,EntrezGene:37856,BIOGRID:63435,DroID:FBgn0034974,DRSC:FBgn0034974,FlyAtlas:CG16786-RA,flyexpress:FBgn0034974,FlyMine:FBgn0034974,GenomeRNAi_gene:37856,modMine:FBgn0034974; derived_computed_cyto=60B8-60B9%3B Limits computationally determined from genome sequence between @P{lacW}Phm<up>k07623</up>@%26@P{lacW}tsr<up>k05633</up>@ and @P{EP}EP503@; gbunit=AE013599; MD5=4a28df05c5f7a49b8fd75a28e3b5759e; length=4429; release=r5.27; species=Dmel; 
CGGATTCGGATTCAGATTCACATTCAGATTCAGATACGTTCGGTTTGGGA
TTCGGATTCATTCGTTGCCACTCCAGCTCTATGCTCCGCGTTGGACCCAC
CGATAGCTTGGCTTTCTGCTACAGTTTCATAATTGTCTCGGCCAGCAGCA
GCGGAGTTCATGATTTCGCTCGGAATATGTTTTAGCCAGATCAGTGCTTG
GAAAATGCACTTTTGAGCGTGTACGTGTATGTGGCAAGTAGCTGGCGAAC
GTGAATGAAAACATGAGCTGCCACTGAACGAAACCCACTCTCGAGCTGGA
AGTGCAAGTGAGTTATCCCGCGGAAGAAAAGAAACTGAATTGATTACCAT
TACCATTCGCGGAGTAGCAGTCTCGGAATTAAATACCAACGACCCAGACA
ATACCGAGCCCAGTTCCAAGCTGGAGGCTCAAGCCTTTCTCTATTCAATG

Do I need to re-import a modified fasta file which has a shorter head information? There seems to be a lot of characters in the header which I cannot understand.

thanks
Abhijit

[Thomas Doktor]

Hi,

You should build a new bowtie index of the Drosophila genome and not of the individual genes as TopHat is designed to align RNA-seq reads against a full genome. This might explain the behaviour of TopHat, although it should have aligned some reads after all. Perhaps the characters in the fasta headers are causing trouble or there are too many contigs for TopHat to handle well.

[gen2prot]

Hi Thomas,

I ran Tophat on the chromosomes and it works wonderfully. I think the fasta header might be the one to blame, since there are characters such as %@><{} etc. It appeared to me as some sort of construct info. Anyways I removed everything except the name of the sequence, and am building the gene index again. Lets see. However, the gene file that I am building the index from is 85 MB in size containing 14964 genes. You think this may cause a problem? Thanks for your help.

Abhijit