What's the longest intron in the human genome?

Question

What's the longest intron in the human genome? Discarding isoforms were a long intron spans more than one exon, i.e. what's the longest distance between two exons without another exons in between.

Answer 1

Short answers: Based on gencode the longest intron for protein coding transcripts, is the intron 3-4 of the transcript ENST00000465127.1 with 1,240,120 bases.

Long answer:

Download the main annotation gff3 file from gencode first. Then run this script:

zcat gencode.v29.annotation.gff3.gz \ 
| awk -v FS="\t" -v OFS="\t" '$3=="exon" && $9 ~ /transcript_type=protein_coding/ {match($0, "transcript_id=([^;]+)", transcript); print $0,transcript[1]}' \
| sort -k1,1V -k10,10 -k4,4g -k5,5g \
| awk -v FS="\t" -v OFS="\t" 'BEGIN {transcript=""; max_d=0; max_line=""} \
 $10 != transcript  {transcript = $10; end = $5; next;} \
 {d=$4-end-1; end=$5; if(d>max_d) {max_d=d; max_line=$0"\t"d;}} \
 END {print max_line}'

It will:

• decompress the file
• filter only lines that describes exons of protein coding transcripts
• append the transcript id at the end of the line
• sort by chromosome, transcript and position
• calculate the distance between two exons of the same transcript
• save the current maximum distance and the annotation line that belongs to it
• print out the line of the exon which follows the longest intron. The last column contains the distance.

The output will be:

chrX    HAVANA  exon    38666121    38666309    .   +   .   ID=exon:ENST00000465127.1:4;Parent=ENST00000465127.1;gene_id=ENSG00000250349.3;transcript_id=ENST00000465127.1;gene_type=protein_coding;gene_name=AF241726.2;transcript_type=protein_coding;transcript_name=AF241726.2-201;exon_number=4;exon_id=ENSE00003644052.1;level=2;protein_id=ENSP00000417050.1;transcript_support_level=5;tag=basic,appris_principal_1,readthrough_transcript;havana_gene=OTTHUMG00000161009.1;havana_transcript=OTTHUMT00000363378.1  ENST00000465127.1   1240120

fin swimmer

EDIT:

gencode also includes transcripts that are only predicted, but never seen, or those where it is not sure that they exists. So you maybe want to just include transcripts having a certain transcript support level. For this you have to change the first awk line. E.g. to include only trasncripts with TSL1 change it to:

awk -v FS="\t" -v OFS="\t" '$3=="exon" && $9 ~ /transcript_type=protein_coding/ && $9 ~ /transcript_support_level=1/ {match($0, "transcript_id=([^;]+)", transcript); print $0,transcript[1]}'

You can also define a range, e.g 1-3:

awk -v FS="\t" -v OFS="\t" '$3=="exon" && $9 ~ /transcript_type=protein_coding/ && $9 ~ /transcript_support_level=[1-3]/ {match($0, "transcript_id=([^;]+)", transcript); print $0,transcript[1]}'

In both cases the result will change to intron 4-5 of ENST00000402655.6 with 1096453 bases.