0 Bracken added reads with GTDB database for all taxonomic units

[alimayy]

Hi there,

Thanks for developing Struo2 and for releasing the GTDB kraken2 dbs prepared with struo2.

I'm using Kraken v2.1.1 and Bracken v2.8 on paired-end animal microbiome data and the struo2 GTDB_release207 kraken2 DB GTDB . The files I have under in my kraken2/bracken DB are as follows with matching md5sums to those in the mentioned DB repository.

-rw-rw-r--  1 ali  ali  287G dec 21 18:53 hash.k2d
-rw-rw-r--  1 ali  ali    64 dec 21 18:53 opts.k2d
-rw-rw-r--  1 ali  ali  9,8M dec 21 18:53 taxo.k2d
-rw-rw-r--  1 ali  ali    59 dec 21 18:53 database150mers.kmer_distrib
-rw-r--r--  1 ali  ali  578M dec 21 18:53 database150mers.kraken

When I run kraken2 and Bracken using this database, I don't get any added reads from Bracken. Any idea why this might be the case, e.g. could I be missing some required files/folders here, or a kraken2 or Bracken version issue?

[nick-youngblut]

Can you please share the log for the run?

[alimayy]

@nick-youngblut , sure

~/packages/kraken2/build/kraken2 --threads 12 \
> --confidence 0 \
> --unclassified-out HAV1804-11-8-18_39-A_S8_unclassified_R#.fastq.gz \
> --output HAV1804-11-8-18_39-A_S8_kraken_output \
> --report HAV1804-11-8-18_39-A_S8_kraken_report.txt \
> --memory-mapping \
> --use-names \
> --db /ramdisk/ \
> --paired \
> --gzip-compressed \
> HAV1804-11-8-18_39-A_S8_L001_R1_001.fastq.gz HAV1804-11-8-18_39-A_S8_L001_R2_001.fastq.gz
Loading database information... done.
9599389 sequences (2882.08 Mbp) processed in 84.199s (6840.5 Kseq/m, 2053.77 Mbp/m).
  6580152 sequences classified (68.55%)
  3019237 sequences unclassified (31.45%)

~/packages/Bracken-2.8/bracken \
> -i HAV1804-11-8-18_39-A_S8_kraken_report.txt \
> -d /ramdisk/ \
> -l S \
> -o HAV1804-11-8-18_39-A_S8_bracken_output_S-lvl.txt \
> -w HAV1804-11-8-18_39-A_S8_bracken_report_S-lvl \
> -r 150 \
> -t 100
 >> Checking for Valid Options...
 >> Running Bracken
      >> python src/est_abundance.py -i HAV1804-11-8-18_39-A_S8_kraken_report.txt -o HAV1804-11-8-18_39-A_S8_bracken_output_S-lvl.txt -k /ramdisk/database150mers.kmer_distrib -l S -t 100
PROGRAM START TIME: 12-22-2022 09:56:34
>> Checking report file: HAV1804-11-8-18_39-A_S8_kraken_report.txt
BRACKEN SUMMARY (Kraken report: HAV1804-11-8-18_39-A_S8_kraken_report.txt)
    >>> Threshold: 100
    >>> Number of species in sample: 51679
          >> Number of species with reads > threshold: 2155
          >> Number of species with reads < threshold: 49524
    >>> Total reads in sample: 9599389
          >> Total reads kept at species level (reads > threshold): 5371999
          >> Total reads discarded (species reads < threshold): 435116
          >> Reads distributed: 0
          >> Reads not distributed (eg. no species above threshold): 773037
          >> Unclassified reads: 3019237
BRACKEN OUTPUT PRODUCED: HAV1804-11-8-18_39-A_S8_bracken_output_S-lvl.txt
PROGRAM END TIME: 12-22-2022 09:56:36
  Bracken complete.

[nick-youngblut]

I don't see anything obvious in the logs that could explain your outcome. Do you get a different result when using the default kraken2 & bracken databases?

[alimayy]

@nick-youngblut yes! Also, I do get added reads with Bracken when I use the struo1 database (with lower % of Kraken2-classified reads, of course).

Can you confirm that you get added reads with Bracken when you use the abovementioned database?