Skip to content

Download

KMCP is implemented in Go programming language, statically-linked executable binary files are freely available.

SIMD instructions support

SIMD extensions including AVX512, AVX2, SSE2 are sequentially detected and used in two packages for better searching performance.

  • pand, for accelerating searching on databases constructed with multiple hash functions.
  • pospop, for batch counting matched k-mers in bloom filters.

ARM architecture is supported, but kmcp search would be slower.

Current Version

v0.9.2 - 2023-05-16 Github Releases (by Release)

  • kmcp profile/cos2simi/filter/index-info/merge-regions/query-fpr:
    • rename/unify the long flag --out-prefix to --out-file.
  • kmcp profile:
    • fix the number of reads belonging to references in the profile when no matches are found, which should be 0 instead of 1.
  • new command:
    • kmcp utils index-density: plotting the element density of bloom filters for an index file. An audience was concerned about it, but the results showed the elements (1s) are uniformly distributed in all BFs.
OS Arch File, 中国镜像 Download Count
Linux 64-bit kmcp_linux_amd64.tar.gz,
中国镜像		 Github Releases (by Asset)
Linux arm64 kmcp_linux_arm64.tar.gz,
中国镜像		 Github Releases (by Asset)
macOS 64-bit kmcp_darwin_amd64.tar.gz,
中国镜像		 Github Releases (by Asset)
macOS arm64 kmcp_darwin_arm64.tar.gz,
中国镜像		 Github Releases (by Asset)
Windows 64-bit kmcp_windows_amd64.exe.tar.gz,
中国镜像		 Github Releases (by Asset)

Notes:

  • please open an issue to request binaries for other platforms or compile from the source.
  • run kmcp version to check update !!!
  • run kmcp autocompletion to update shell autocompletion script !!!

Installation

Method 1: Install using conda Anaconda Cloud downloads

conda install -c bioconda kmcp

Method 2: Download binaries

Download the compressed executable file of your operating system, and decompress it with tar -zxvf *.tar.gz command or other tools. And then:

  • For Linux-like systems

    • If you have root privilege, simply copy it to /usr/local/bin:

      sudo cp kmcp /usr/local/bin/

    • Or copy to anywhere in the environment variable PATH:

      mkdir -p $HOME/bin/; cp kmcp $HOME/bin/

  • For Windows, just copy kmcp.exe to C:\WINDOWS\system32.

Method 3: Compile from source

  1. Install go

    wget https://go.dev/dl/go1.17.13.linux-amd64.tar.gz

tar -zxf go1.17.13.linux-amd64.tar.gz -C $HOME/

# or 
#   echo "export PATH=$PATH:$HOME/go/bin" >> ~/.bashrc
#   source ~/.bashrc
export PATH=$PATH:$HOME/go/bin

  2. Compile KMCP

    # ------------- the latest stable version -------------

go get -v -u github.com/shenwei356/kmcp/kmcp

# The executable binary file is located in:
#   ~/go/bin/kmcp
# You can also move it to anywhere in the $PATH
mkdir -p $HOME/bin
cp ~/go/bin/kmcp $HOME/bin/

# --------------- the development version --------------

git clone https://github.com/shenwei356/kmcp
cd kmcp/kmcp/
go build

# The executable binary file is located in:
#   ./kmcp
# You can also move it to anywhere in the $PATH
mkdir -p $HOME/bin
cp ./kmcp $HOME/bin/

Shell-completion

Supported shell: bash|zsh|fish|powershell

Bash:

# generate completion shell
kmcp autocompletion --shell bash

# configure if never did.
# install bash-completion if the "complete" command is not found.
echo "for bcfile in ~/.bash_completion.d/* ; do source \$bcfile; done" >> ~/.bash_completion
echo "source ~/.bash_completion" >> ~/.bashrc

Zsh:

# generate completion shell
kmcp autocompletion --shell zsh --file ~/.zfunc/_kmcp

# configure if never did
echo 'fpath=( ~/.zfunc "${fpath[@]}" )' >> ~/.zshrc
echo "autoload -U compinit; compinit" >> ~/.zshrc

fish:

kmcp autocompletion --shell fish --file ~/.config/fish/completions/kmcp.fish

Release History

v0.9.1 - 2022-12-26 Github Releases (by Release)

  • kmcp search
    • faster speed for ARM architectures.
    • fix compilation for ARM architectures.

v0.9.0 - 2022-09-28 Github Releases (by Release)

  • compute:
    • smaller output files and faster speed.
    • more even genome splitting.
  • index:
    • faster speed due to smaller input files.
  • search:
    • more accurate and smaller query FPR following Theorem 2 in SBT paper, instead of the Chernoff bound.
    • change the default value of -f/--max-fpr from 0.05 to 0.01.
    • 10-20% speedup.
  • profile:
    • more accurate abundance estimation using EM algorithm.
    • change the default value of -f/--max-fpr from 0.05 to 0.01.
    • mode 0: change the default value of -H/--min-hic-ureads-qcov from 0.55 to 0.7.
    • increase float width of reference coverage in KMCP profile format from 2 to 6.
  • util query-fpr:
    • compute query FPR following Theorem 2 in SBT paper, instead of the Chernoff bound.
  • new commands:
    • utils split-genomes for splitting genomes into chunks.
    • utils ref-info for printing information of reference (chunks), including the number of k-mers and the actual false-positive rate.

v0.8.3 - 2022-08-15 Github Releases (by Release)

  • kmcp: fix compiling from source for ARM architectures.#17
  • search:
    • fix searching with paired-end reads where the read2 is shorter than the value of --min-query-len. #10
    • fix the log. #8
    • a new flag -f/--max-fpr: maximum false positive rate of a query (default 0.05). It reduces the unnecessary output when searching with a low minimum query coverage (-t/--min-query-cov).
  • profile:
    • recommend using the flag --no-amb-corr to disable ambiguous reads correction when >= 1000 candidates are detected.
    • fix logging when using --level strain and no taxonomy given.

v0.8.2 - 2022-03-26 Github Releases (by Release)

  • search:
    • flag -g/--query-whole-file:
      • fix panic for invalid input.
      • add gaps of k-1 bp before concatatenating seqs.
    • add warning for invalid input.
  • profile:
    • allow modifying parts of parameters in preset profiling modes. #5
    • decrease thresholds of minimum reads and unique reads in preset profiling modes 1 and 2 for low coverage sequence data. the profiling results generated with mode 3 in the manuscript are not affected.

v0.8.1 - 2022-03-07 Github Releases (by Release)

  • update help message, show common usages, add examples, add notes to important options.

v0.8.0 - 2022-02-24 Github Releases (by Release)

  • commands:
    • new command utils cov2simi: Convert k-mer coverage to sequence similarity.
    • new command utils query-fpr: Compute the maximum false positive rate of a query.
  • compute:
    • update doc.
    • add flags compatibility check.
  • search:
    • output the false positive rate of each match, rather than the FPR upper bound of the query. this could save some short queries with high similarity.
    • change default values of reads filter, because clinical data contain many short reads.
      • -c/--min-uniq-reads: 30 -> 10.
      • -m/--min-query-len: 70 -> 30.
    • update doc.
  • profile:
    • rename flags:
      • --keep-main-matches -> --keep-main-matches.
      • --keep-perfect-match -> --keep-perfect-matches.
    • change default values:
      • --max-qcov-gap: 0.2 -> 0.4.
    • mode 0 (pathogen detection):
      • switch on flag --keep-main-matches
      • use --max-qcov-gap 0.4
    • update doc.

v0.7.1 - 2022-02-08 Github Releases (by Release)

  • profile:
    • new flag --metaphlan-report-version and the default value is 3. #4
    • column name renamed: from fragsFrac, fragsRelDepth, fragsRelDepthStd to chunksFrac, chunksRelDepth, chunksRelDepthStd.
    • fix computation of chunksRelDepth.
    • slightly improve sensitivity for -m 0.

v0.7.0 - 2022-01-24 Github Releases (by Release)

  • commands:
    • new command utils filter: Filter search results and find species-specific queries.
    • new command utils merge-regions: Merge species/assembly-specific regions.
    • rename info to utils index-info.
  • compute:
    • skip k-mer containing Ns.
    • when splitting genome into fragments, sequences are concatenated with k-1 'N's instead of directly concatenation. It eliminates fake k-mers at the concatenation position.
    • set default value for flag -N/--ref-name-regexp: (?i)(.+)\.(f[aq](st[aq])?|fna)(.gz)?$.
    • fix a rare bug when splitting FASTQ files.
  • search:
    • support searching with paired-end reads which has a higher specificity and a lower sensitivity. A flag --try-se is added for search read1/read2 when the paired end reads have no hits.
    • fix matches order of a query.
    • fix queries with many Ns.
    • change default value of flag -t/--min-query-qcov from 0.6 to 0.55 (similarity ~96.5%).
    • change default value of flag -n/--keep-top-scores from 5 to 0, i.e., keep all matches by default.
    • new flag -w/--load-whole-db: load all index files into memory.
    • 10-25% faster.
    • better log.
  • merge:
    • fix adding up hits.
    • fix bug of incorrect order, reduce memory usage.
    • support one input file.
  • profile:
    • change analysis workflow, using 4 stages.
    • output format change: new column coverage, fragsRelDepth and fragsRelDepthStd.
    • change default file extension of binning file.
    • check if the taxid of a target is given by taxid mapping file.
    • automatically switch to the new taxid for a merged one.
    • change computation of score.
    • new flag -d/--max-frags-depth-stdev.
    • new option -m/--mode.
    • change default value of flag -t/--min-query-qcov from 0.6 to 0.55 (similarity ~96.5%).
    • change default value of flag -n/--keep-top-qcovs from 5 to 0 (keep all matches).
    • change default value of falg -f/--max-fpr from 0.01 to 0.05.
    • change default value of flag -H/--min-hic-ureads-qcov from 0.8 to 0.75 (similarity ~98%).
    • faster search result parsing.

v0.6.0 - 2021-08-13

  • new command:
    • merge: merge search results from multiple databases.
  • compute:
    • fix splitting very short genomes.
    • remove flag -e/--exact-number, making it default.
  • index:
    • do not roundup sizes of indexes. The searching speed is not affected and even faster due to optimization of search command.
    • use three k-mers thresholds to control index file size.
    • better control of cocurrency number and better progress bar.
    • do not support RAMBO index anymore.
  • search:
    • 1.37X speedup, and faster for database with two or more hash functions.
    • new flag -S/--do-not-sort.
  • profile:
    • fix a nil pointer bug when no taxid mapping data given.
    • fix number of ureads.
    • new flag -m/--keep-main-matches and --max-score-gap

v0.5.0 - 2021-06-24

  • compute:
    • support multiple sizes of k-mer.
    • fix bug of --by-seq.
    • more log.
  • index:
    • default block size is computed by -j/--threads instead of number of CPUs.
  • search:
    • show real-time processing speed.
    • new flag -g/--query-whole-file.
    • new flag -u/--kmer-dedup-threshold.
    • new flag -m/--min-query-len.
    • increase speed for database with mulitple hashes.
  • profile:
    • better decision of the existence of a reference.
    • new flag -B/--binning-result for output reads binning result.
    • new flag -m/--norm-abund.

v0.4.0 - 2021-04-08

  • new command:
    • profile for generating taxonomic profile from search result.
  • compute:
    • new flag -B/--seq-name-filter for filtering out unwanted sequences like plasmid.
    • new flag -N/--ref-name-regexp for extracting reference name from sequence file.
  • search:
    • change default threshold value.
    • new flag -n/--keep-top-scores for keeping matches with the top N score.

v0.3.0 - 2021-03-16

  • use --quiet to replace --verbose, making printing log info default.
  • search:
    • fix computing intersetion between repeats.
    • fix closing mmap on Windows.
    • change output format and add Jaccard Index.
    • speedup by parallelizing name mapping and database closing.
    • flush result immediately.
    • keep the output order by default
  • compute: change default file regexp for matching .fna files.
  • autocompletion: support bash, zsh, fish, powershell.

v0.2.1 - 2020-12-31

  • index: reduce memory occupation.

v0.2.0 - 2020-12-30

  • Add support of RAMBO like indexing.
  • Limit to only one input database.
  • Change output format.

v0.1.0 - 2020-xx-xx

  • First release with basic function.