Listing which biotypes are present in an Ensembl .gtf file

[heso]

I'm trying to find out which gene_biotypes are described in the Human gtf file from Ensembl.
The head of the file looks like that:

1       havana  gene    19072110        19075511        .       -       .       gene_id "ENSG00000272084"; gene_version "1"; gene_name "RP5-1126H10.2"; gene_source "havana"; gene_biotype "3prime_overlapping_ncRNA"; havana_gene "OTTHUMG00000185684"; havana_gene_version "1";
1       havana  transcript      19072110        19075511        .       -       .       gene_id "ENSG00000272084"; gene_version "1"; transcript_id "ENST00000606379"; transcript_version "1"; gene_name "RP5-1126H10.2"; gene_source "havana"; gene_biotype "3prime_overlapping_ncRNA"; havana_gene "OTTHUMG00000185684"; havana_gene_version "1"; transcript_name "RP5-1126H10.2-001"; transcript_source "havana"; transcript_biotype "3prime_overlapping_ncRNA"; havana_transcript "OTTHUMT00000470990"; havana_transcript_version "1"; tag "basic"; transcript_support_level "NA";
1       havana  exon    19072110        19075511        .       -       .       gene_id "ENSG00000272084"; gene_version "1"; transcript_id "ENST00000606379"; transcript_version "1"; exon_number "1"; gene_name "RP5-1126H10.2"; gene_source "havana"; gene_biotype "3prime_overlapping_ncRNA"; havana_gene "OTTHUMG00000185684"; havana_gene_version "1"; transcript_name "RP5-1126H10.2-001"; transcript_source "havana"; transcript_biotype "3prime_overlapping_ncRNA"; havana_transcript "OTTHUMT00000470990"; havana_transcript_version "1"; exon_id "ENSE00003701142"; exon_version "1"; tag "basic"; transcript_support_level "NA";
1       havana  gene    201464383       201465146       .       -       .       gene_id "ENSG00000224818"; gene_version "1"; gene_name "RP11-134G8.10"; gene_source "havana"; gene_biotype "3prime_overlapping_ncRNA"; havana_gene "OTTHUMG00000189503"; havana_gene_version "1";
1       havana  transcript      201464383       201465146       .       -       .       gene_id "ENSG00000224818"; gene_version "1"; transcript_id "ENST00000430471"; transcript_version "1"; gene_name "RP11-134G8.10"; gene_source "havana"; gene_biotype "3prime_overlapping_ncRNA"; havana_gene "OTTHUMG00000189503"; havana_gene_version "1"; transcript_name "RP11-134G8.10-001"; transcript_source "havana"; transcript_biotype "3prime_overlapping_ncRNA"; havana_transcript "OTTHUMT00000479807"; havana_transcript_version "1"; tag "basic"; transcript_support_level "5";
1       havana  exon    201465061       201465146       .       -       .       gene_id "ENSG00000224818"; gene_version "1"; transcript_id "ENST00000430471"; transcript_version "1"; exon_number "1"; gene_name "RP11-134G8.10"; gene_source "havana"; gene_biotype "3prime_overlapping_ncRNA"; havana_gene "OTTHUMG00000189503"; havana_gene_version "1"; transcript_name "RP11-134G8.10-001"; transcript_source "havana"; transcript_biotype "3prime_overlapping_ncRNA"; havana_transcript "OTTHUMT00000479807"; havana_transcript_version "1"; exon_id "ENSE00001739739"; exon_version "1"; tag "basic"; transcript_support_level "5";

I tried

cat Homo_sapiens.GRCh38.85.gtf| cut -f 9| awk -F ';' '{print $7}'| sort | uniq

...but since the gene_biotype is not always in $7 I also get entries like:

gene_source "ensembl"
gene_source "ensembl_havana"
gene_source "havana"

...which makes me suspect that I might miss some biotypes

Can somebody please help me to improve the code to retrieve the full list of gene_biotype's

Thank you in advance....

[wdecoster]

I would do it like this:

cat Homo_sapiens.GRCh38.85.gtf | cut -f 9 | tr ';' '\n' | grep "gene_biotype" | sed 's/gene_biotype//' | sort -u