What does the 'E%' represent in BUSCO results?

Question

[林明德]

I performed a BUSCO genome mode analysis using the insecta_odb10 dataset.

The results were as follows: C: 99.8% [S: 91.1%, D: 8.7%], F: 0.1%, M: 0.1%, n: 1367, E: 3.7%.

Can someone help me interpret what "E: 3.7%" stands for?

I would really appreciate any help.

BUSCO Image tag version: ezlabgva/busco:v5.7.0_cv1 and ezlabgva/busco:v5.7.1_cv1

Code: nohup busco -i genome.fasta --auto-lineage-euk -o /path/to/outdir -m genome -c 80 &

[GenoMax]

According to ChatGPT:

In BUSCO, the "E" category represents "End" or "Endof" gene fragments. These are orthologous groups for which the gene model is fragmented at the end of the sequence, indicating that the sequence is incomplete at the 3' or 5' end.

[Philipp Bayer]

funny enough i tried the same thing (i've used BUSCO heaps times but never saw the E:! even the manual lists 'C:89.0%[S:85.8%,D:3.2%],F:6.9%,M:4.1%,n:3023', without the E:) and ChatGPT waffled something about E: being the percentage of eukaryotic genes in the proteome. One of us is wrong, or both :)

Answer 1

[Philipp Bayer]

Ah your edit solved this. 50 BUSCOs contain internal stop codons, indicating they might be erroneous gene models, or the genes themselves are inactive/lost.

50 / 1364 * 100 = 3.66 ~ 3.7%. That's a new feature! So ChatGPT was wrong in both comments :)

This seems to be a new feature, BUSCO 5.6.0, the commit that introduces this detail is 4 months old. https://gitlab.com/ezlab/busco/-/blame/master/src/busco/busco_tools/hmmer.py?page=2#L1158

[林明德]

Thank you for your answer! It really helped me resolve a doubt!

[Philipp Bayer]

No worries! I found the commit with the maths too: https://gitlab.com/ezlab/busco/-/blame/master/src/busco/busco_tools/hmmer.py?page=2#L1350

                100 * self.complete_stop_codon_count / complete_count