rocksdb fastmultigather query result sums to > 1

[bluegenes]

Fastmultigather error

I have some protein fastmultigather results from running against rocksdb, and am getting a taxonomy error,

ERROR: Summarized fraction is > 100% of the query! This should not be possible. Please check that your input files come directly from a single gather run per query.

I tracked down the entry, and when I go to look at the fastmultigather results file, I find two results that do indeed add to > 100%
f_unique_to_query: 0.880 + 0.238 = 1.118

the relevant lines from the fastmultigather output:

grep "AB087737 Macacine alphaherpesvirus 1" output.spillover/fastmultigather/spillover-x-vmr_MSL38_v1.protein.k10-sc2.t8.fmgather.csv

intersect_bp,f_orig_query,f_match,f_unique_to_query,f_unique_weighted,average_abund,median_abund,std_abund,match_filename,match_name,match_md5,f_match_orig,unique_intersect_bp,gather_result_rank,remaining_bp,query_filename,query_name,query_md5,query_bp,ksize,moltype,scaled,query_n_hashes,query_abundance,query_containment_ani,match_containment_ani,average_containment_ani,max_containment_ani,n_unique_weighted_found,sum_weighted_found,total_weighted_hashes
618,0.8803418803418803,0.016539985012311315,0.8803418803418803,0.8803418803418803,1.0,1.0,0.0,genbank/proteomes/GCF_000844145.1_protein.faa.gz,GCF_000844145.1 Macacine alphaherpesvirus 1,a026087a999d322345466b16e4bd6c76,0.016539985012311315,618,0,84,output.spillover/protein/AB087737.faa.gz,AB087737 Macacine alphaherpesvirus 1,5e19143eaf4bc341d85a97eb193cc077,702,10,protein,1,702,false,0.995760845802698,0.8722035240343783,0.9339821849185381,0.995760845802698,0,0,702
583,0.8304843304843305,0.0005351743330389875,0.23809523809523808,0.23809523809523808,1.0,1.0,0.0,genbank/proteomes/GCF_018580865.1_protein.faa.gz,GCF_018580865.1 Macacine alphaherpesvirus 1,f20b4ebd52ca4881b4ab8566be278738,0.015600331808086483,20,1,64,output.spillover/protein/AB087737.faa.gz,AB087737 Macacine alphaherpesvirus 1,5e19143eaf4bc341d85a97eb193cc077,702,10,protein,1,702,false,0.9532899700399211,0.7779466993890651,0.8656183347144931,0.9532899700399211,0,0,702

When I check the original input file, this query is only present once, so I don't think we searched it twice and tried to combine the results.

grep "AB087737 Macacine alphaherpesvirus 1" output.spillover/spillover.fromfile.csv
AB087737 Macacine alphaherpesvirus 1,output.spillover/genomic/AB087737.fna.gz,output.spillover/protein/AB087737.faa.gz

[ctb]

I'm happy to dive into this, but more info would be really helpful ;).

• either a small reproducible test case, or (at worst) the original example;
• and/or, the commands used, along with the location of the input and output files;

[bluegenes]

I grepped out the signature associated with this identifier to use as test data. Data on farm in /home/ntpierce/2023-fmg-bug

Query file:

• AB087737.protein-k10-sc2.zip - scaled 2

ICTV-VMR database file(s):

• vmr_MSL38_v1.protein-k10-sc2.rocksdb - protein k10, scaled 2 rocksdb database
• vmr_MSL38_v1.protein.zip - protein k7,k10, scaled 1 zip database

AB087737 test outputs:

fastgather output:

• AB087737.gather.csv
• AB087737.prefetch.csv

gather output:

• AB087737.protein-k10-sc2.gather.csv

full queryset rocksdb fastmultigather output files:

• spillover-x-vmr_MSL38_v1.protein.k10-sc1.t4.fmgather.csv
• spillover-x-vmr_MSL38_v1.protein.k10-sc2.t8.fmgather.csv
• spillover-x-vmr_MSL38_v1.protein.k10-sc2.t8.fmgather.with-lineages.csv