BUSCO results apparently inconsistent

Question

I have two assemblies obtained from two different softwares for the same run data. In one of them, I get a BUSCO score of 69,4%. In the other one, I get 68,5%. The first assembly covers a 99,7% fraction of the reference while the second one covers 100%.

The relevant thing is that the reference genome also gets 68,5% (less than the first assembly). How can I interpret this situation?

I identify these options:

• The first assembly is showing a different strain.
• BUSCO is imprecise
• The first assembly is merging DNA from another organism
• The first assembly is a misassembly, repeating part of the sequence and BUSCO interprets that as new genes
• The sample really is an organism that has more genes than the reference.

The organism is a bacteria and I am using bacteria_odb10 as the lineage dataset.

This is the BUSCO result for the reference (LN554883.1 Chlamydophila abortus strain 1H):

C:68.5%[S:68.5%,D:0.0%],F:8.1%,M:23.4%,n:124       
85  Complete BUSCOs (C)            
85  Complete and single-copy BUSCOs (S)    
0   Complete and duplicated BUSCOs (D)     
10  Fragmented BUSCOs (F)              
29  Missing BUSCOs (M)             
124 Total BUSCO groups searched

This for the first assembly:

C:69.4%[S:69.4%,D:0.0%],F:8.9%,M:21.7%,n:124       
86  Complete BUSCOs (C)            
86  Complete and single-copy BUSCOs (S)    
0   Complete and duplicated BUSCOs (D)     
11  Fragmented BUSCOs (F)              
27  Missing BUSCOs (M)             
124 Total BUSCO groups searched     

And this for the second assembly:

C:68.5%[S:0.0%,D:68.5%],F:8.1%,M:23.4%,n:124       
85  Complete BUSCOs (C)            
0   Complete and single-copy BUSCOs (S)    
85  Complete and duplicated BUSCOs (D)     
10  Fragmented BUSCOs (F)              
29  Missing BUSCOs (M)             
124 Total BUSCO groups searched

I see now that the first assembly also contains a lot of contigs with human DNA. I guess BUSCO finds a match for some genes in the human DNA?

Answer 1

The core idea behind BUSCO is that there are a number of genes that are Unique Single-Copy Orthologues (USCOs) that should be present in most genomes and should only appear once in the genome. This is only a guess, however, and not a universal rule.

Sometimes genomes can have multiple copies of these USCOs, and these duplicated genes reduce the overall score (which is based on the proportion of complete single-copy genes that are present as a fraction of the entire USCO set).

Regarding human DNA, bacteria are a curiously diverse group, with genomes that have more similarities in function than in form. Given this diversity, it wouldn't surprise me if contaminating DNA could increase the BUSCO score.