Converting fasta to gtf using reference genome

Question

[KI]

Hello,

I am trying to convert fasta file of the transcripts to gtf using a reference genome, but I am having some trouble. I specifically need gtf file of mRNA to run a program. The genome and the transcript data were downloaded from (https://parasite.wormbase.org/ftp.html). Following are the approaches I have taken so far:

1. fasta-to-gtf script (https://github.com/willblev/fasta-to-gtf)

I found a script converts fasta files to gtf using a reference genome file, but I keep getting a syntax error below. I checked the script and could not identify or manipulate the syntax problem in line 3.

./fasta-to-gtf.py: line 3: syntax error near unexpected token `('
./fasta-to-gtf.py: line 3: `def usage():'

1. Alignment using hisat2

I tried to align the transcript file to the genome using hisat2. However, I'm not sure how to run hisat2 with the transcript file. I tried to use the transcript.fasta and genome.fasta as mate 1 and mate 2 with the index generated from annotation file, but I received the error below. The only fix I found was on bowtie2 github that suggests redownloading the latest version of the program, but I am already using the latest version of hisat2 (https://github.com/BenLangmead/bowtie2/issues/149).

Segmentation fault (core dumped)
(ERR): hisat2-align exited with value 139

Please let me know if you have any other suggestions on this conversion process and if you need any further information.

Thank you!

[sangram_keshari]

If you not sure how to run hisat2 -> this nature protocol paper can be a good read.

Answer 1

[h.mon]

First of all, why the GFF3 / GTF files provided at WormBase Parasite are not suitable? Why do you need to recreate information which is already available?

As for the programs you tried: the fasta-to-gtf.py script don't have a shebang line, so you can't make it executable and run it directly. You have to cal it with python fasta-to-gtf.py:

python fasta_to_gtf.py -h

Outputs:

~~~~~~~~~~~~~~~~~~~~~~~~~~
~~~~~~~fasta-to-gtf~~~~~~~
~~~~~~~~~~~~~~~~~~~~~~~~~~
fasta-to-gtf.py -g <path/reference_genome.fasta> -a <path/assembly.fasta> -o <path/output_name.gtf>
This is a python script which produces GTF annotations using a transcriptome assembly and reference genome (in fasta format).
The script requires python 2.7 (or later) and the following python modules:
getopt, CGAT
To run this script, you must specify a reference genome, an assembly, and an output file name.
Use -v or --verbose to print intermediate status updates

As for HISAT2:

I tried to use the transcript.fasta and genome.fasta as mate 1 and mate 2 with the index generated from annotation file,

It is not clear what you did here, but it seems to me you tried HISAT2 with -1 transcript.fasta -2 genome.fasta. Is my interpretation correct? If so, you are running it incorrectly. You have to build the index with the genome.fasta, then you align the transcript.fasta as single-end unpaired reads with -U transcripts.fasta. And HISAT2 output will be a SAM alignment file, not GTF / GFF3, you would need additional steps to arrive at a GTF / GFF3 file.

But anyway, why don't you use the GTF / GFF3 provided at the WormBase Parasite site?