Create a fasta and gtf file

Question

[ATRX]

Hi,

Is there a tool to create a fasta file using the long gene length sequence (i.e., TSE to TSS) that includes all the genomic elements, including 5' UTR, exons, introns 3'UTR from hg38? And then create a new gtf file using the fasta file subset and original gtf file information?

Thank you!

Answer 1

[Juke34]

You may use AGAT.

See here for agat_sp_extract_sequences.pl

Command: agat_sp_extract_sequences.pl -g infile.gff -f infile.fasta -t gene

[ATRX]

This is awesome. Thank you and it worked like a charm!

[ATRX]

Juke34 I have another question. Is there a way to extract the longest gene sequence without the introns? For example:

agat_sp_extract_sequences.pl -g infile.gff -f infile.fasta -t exon, but I want the sequence in the gene level and not in the transcript level. Thanks a lot!

[Juke34]

Transcript level and gene level is the same when you have only one isoform. If you have several isoforms you may filter them to keep the longest with agat, or merge locations with bedtools intersect

Answer 2

[michael.ante]

Hi, You can use bedtools' getfasta function to extract a sequence from a fasta file, creating a new fasta. Since you work with human, you should have the complete fasta and gene information at hand. Just create a bed file (I guess gtf works as well) with only start and end of your long gene.

bedtools getfasta -fi hg36.fasta -bed myGene.bed -fo myGene.fasta

Then take the original gtf and reduce all coordinates by the one from your bed file (check for 0-based vs. 1-based coordinates in bed vs. gtf ). Adjust the contig name if necessary and load everything into IGV and check if it fits.

[ATRX]

Thank you! This is very helpful.