Failure to generate index in sr mode with 2.18

[wwood]

Hi,

I updated to 2.18 and rerun my test code for my tool which uses minimap2 under its hood. There was a new test failure, which seems to originate from minimap2 itself. Here's code to reproduce just with minimap2 installed and a small dataset to reproduce. Indexing works by default, but fails when -x sr is specified it fails:

(/home/ben/e/coverm-dev) b2:20210421:~/git/coverm-local$ wget https://raw.githubusercontent.com/wwood/CoverM/main/tests/data/7seqs.fna
--2021-04-21 11:53:16--  https://raw.githubusercontent.com/wwood/CoverM/main/tests/data/7seqs.fna
Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.109.133, 185.199.108.133, 185.199.111.133, ...
Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.109.133|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 57241 (56K) [text/plain]
Saving to: '7seqs.fna’

7seqs.fna                                            100%[=====================================================================================================================>]  55.90K  --.-KB/s    in 0.001s  

2021-04-21 11:53:16 (58.9 MB/s) - '7seqs.fna’ saved [57241/57241]

(/home/ben/e/coverm-dev) b2:20210421:~/git/coverm-local$ minimap2 -x sr -d /tmp/my.mmi 7seqs.fna 
[M::mm_idx_gen::0.002*1.41] collected minimizers
[M::mm_idx_gen::0.004*1.88] sorted minimizers
[M::main::0.005*1.65] loaded/built the index for 7 target sequence(s)
[M::mm_idx_stat] kmer size: 21; skip: 11; is_hpc: 0; #seq: 7
[M::mm_idx_stat::0.005*1.60] distinct minimizers: 9434 (100.00% are singletons); average occurrences: 1.000; average spacing: 6.044; total length: 57016
ERROR: failed to map the query file

(/home/ben/e/coverm-dev) b2:20210421:~/git/coverm-local$ minimap2 -d /tmp/my.mmi 7seqs.fna 
[M::mm_idx_gen::0.002*1.46] collected minimizers
[M::mm_idx_gen::0.003*1.91] sorted minimizers
[M::main::0.004*1.66] loaded/built the index for 7 target sequence(s)
[M::mm_idx_stat] kmer size: 15; skip: 10; is_hpc: 0; #seq: 7
[M::mm_idx_stat::0.004*1.62] distinct minimizers: 10584 (99.98% are singletons); average occurrences: 1.000; average spacing: 5.386; total length: 57016
[M::main] Version: 2.18-r1015
[M::main] CMD: minimap2 -d /tmp/my.mmi 7seqs.fna
[M::main] Real time: 0.005 sec; CPU: 0.008 sec; Peak RSS: 0.005 GB

(/home/ben/e/coverm-dev) b2:20210421:~/git/coverm-local$ minimap2 --version
2.18-r1015

Any ideas? Thanks in advance.

[alienzj]

@wwood Your case is indeed reproducible. I have encountered a similar problem recently.

[lh3]

Thanks. That was a bug lurking around for some time.

[wwood]

No problem, thanks for the fix. I guess it will be out in the next release after 2.18? Until then I'll pin my downstream tool to 2.17.