I am visualizing blastn results from a metagenomic sample using Megan 5. Blastn output was filtered using a Customed ruby script as reportedin the paper https://genomebiology.biomedcentral.com/articles/10.1186/s13059-017-1299-7:
We downloaded and ran MEGAN (v5.10.6) from http://ab.inf.uni-tuebingen.de/software/megan5/. We ran MEGAN in non-interactive (command line) mode as follows:
$ MEGAN/tools/blast2lca --format BlastTAB –topPercent 10 \
--input < sample > .blast --output < sample > _read_assignments.txt
This MEGAN command returns the lowest common ancestor (LCA) taxon in the NCBI Taxonomy for each read. The topPercent option (default value 10) discards any hit with a bitscore less than 10% of the best hit for that read.
We used a custom Ruby script, summarize_megan_taxonomy_file.rb, to sum the per-read assignments into cumulative sums for each taxon. The script enforced the MEGAN parameter, Min Support Percent = 0.1, which requires that at least this many reads (as a percent of the total reads with hits) be assigned to a taxon for it to be reported. Taxa with fewer reads are assigned to the parent in the hierarchy. Output files were given the suffix “BlastMeganFiltered” to indicate that an abundance threshold (also called a filter in this manuscript) was applied. We produced a second set of output files using 0.01 as the minimum percentage and named with the suffix “BlastMeganFilteredLiberal.”
I would like to undertand wht the bit disjoint score is to understand if filtering done by the used script is too stringet and wether to modify it. thank you
