Ilya Shlyakhter

the -hm option of filter isn't in the pdf doc

add option to count each kmer at most once per read

1

Currently, a kmer that occurs twice in the same read gets a count of 2, same as a kmer that occurs once in two different reads. But the latter kmer...

Sometimes a kmer might appear in different reads, so it looks real, but all these reads are duplicates. It would be good if there was an option to drop kmers...

feature request: support IUPAC ambiguity codes

8

When a kmer in a read includes IUPAC ambiguity codes, add all possible concretizations of that kmer.

compressed output file support for filtering reads

1

The output file format for filtering of reads is said to default to "same as input". But if the input is myfile.fasta.gz , the output is uncompressed even if named...

When filtering reads, it would help to have "just the passed read names" as one output format, for use with https://broadinstitute.github.io/picard/command-line-overview.html#FilterSamReads

For using with makefile/pipeline tools like snakemake, it would help to return a non-zero error code in case of failure.

E.g. changing kmc1_db_writer.h does not cause kmc_tools binary to be rebuilt.

filter command doesn't support multi-line fastas

4

@marekkokot The filter command does not seem to handle correctly fasta files where the sequence is split across multiple lines. Also, if the file is too large, it fails with...

feature request: omit kmers with more than x bases below given quality

1

Add option to ignore untrustworthy kmers in reads, with "untrustworthy" defined as "having more than this many bases below a given quality".