KING Download History
Linux-king227.tar.gz For GNU/LINUX 64 bit systems
Mac-king227.tar.gz For Mac Workstations
Windows-king227.zip For Windows Workstations
KING2.2.7code.tar.gz C++ source code
Major changes in KING 2.2.7 include:
1. The same bug introduced in version 2.2.5 for --ibdseg is now completly fixed.
2. A few other minor improvements, including the new KING website "kingrelatedness.com" when king is typed.
Linux-king226.tar.gz For GNU/LINUX 64 bit systems
Mac-king226.tar.gz For Mac Workstations
Windows-king226.zip For Windows Workstations
KING2.2.6code.tar.gz C++ source code
Major changes in KING 2.2.6 include:
1. A bug in --ibdseg that caused a small proportion of datasets not able to finish running in version 2.2.5 is now fixed
2. --pca --projection --rplot can be used for ancestry inference
3. A number of minor issues are fixed. Sorry for the long delay of this release, partially because of the pandemics.
Linux-king225.tar.gz For GNU/LINUX 64 bit systems
Mac-king225.tar.gz For Mac Workstations
Windows-king225.zip For Windows Workstations
KING2.2.5code.tar.gz C++ source code
Major changes in KING 2.2.5 include:
1. A bug in --unrelated is fixed.
2. --ibdseg is substantially improved.
3. --build is improved.
4. A number of minor issues are fixed.
Linux-king224.tar.gz For GNU/LINUX 64 bit systems
Mac-king224.tar.gz For Mac Workstations
Windows-king224.zip For Windows Workstations
KING2.2.4code.tar.gz C++ source code
Major changes in KING 2.2.4 include:
1. All population structure analyses are substantially faster than before.
2. KING is now the tool of choice for Linux users to infer population structure and ancestry.
3. --pcs is newly added to allow the specification of the number of PCs with a default value of 10.
Linux-king223.tar.gz For GNU/LINUX 64 bit systems
Mac-king223.tar.gz For Mac Workstations
Windows-king223.zip For Windows Workstations
KING2.2.3code.tar.gz C++ source code
Major changes in KING 2.2.3 include:
1. The precompiled Linux version of KING now works for all Linux 64 bit systems (e.g., CentOS 7 and Ubuntu)
2. The precompiled Linux version of KING can now infer population structure (--mds, --pca) properly (by using static LAPACK)
3. Precompiled binaries are now available for Windows machines (for most analyses)
4. IBD segment analysis and Run of Homozygosity anaysis now apply to chromosome X as well
5. A bug is fixed for --related analysis when the number of SNPs is less than 4096
Linux-king222.tar.gz For GNU/LINUX 64 bit systems
Ubuntu-king222.tar.gz For Ubuntu 64 bit systems
Mac-king222.tar.gz For Mac Workstations
KING2.2.2code.tar.gz C++ source code
There is only one change in KING 2.2.2. A bug introduced in version 2.2.1 is now fixed.
This bug occurs when reading in multiple datasets with identical SNP sets, such as in this example.
Linux-king221.tar.gz For GNU/LINUX 64 bit systems
Ubuntu-king221.tar.gz For Ubuntu 64 bit systems
Mac-king221.tar.gz For Mac Workstations
KING2.2.1code.tar.gz C++ source code
Major changes in KING 2.2.1 include:
1. --rplot is improved for multiple KING analyses. See KING Tutorial: Visualization of Families.
e.g., --roh --rplot generates run of homozygosity plots for each inbred individual
2. A minor bug in --ibdseg and --roh is now fixed (regarding maxIBD1 and maxIBD2, not affecting the main inference)
3. A bug in --build is now fixed
4. -b KG.bed,ex.bed --mds --proj --rplot can infer ancestry of each ex individual using the KG samples as reference
Linux-king22.tar.gz For GNU/LINUX 64 bit systems
Ubuntu-king22.tar.gz For Ubuntu 64 bit systems
Mac-king22.tar.gz For Mac Workstations
KING2.2code.tar.gz C++ source code
Major changes in KING 2.2 include:
1. Visualization of families is now available. See KING Tutorial: Visualization of Families.
2. Multiple datasets as input is even more flexible than before.
Now duplicate IDs are allowed between the first dataset (Reference) and other datasets.
3. A bug in --ibdseg --degree 2 is now fixed (which only exists in version 2.1.8)
We believe the current version of KING is rather integrated and relatively stable, and we decided to change the version number to 2.2 (instead of 2.1.9).
Linux-king218.tar.gz For GNU/LINUX 64 bit systems
Ubuntu-king218.tar.gz For Ubuntu 64 bit systems
Mac-king218.tar.gz For Mac Workstations
KING2.1.8code.tar.gz C++ source code
Major changes in KING 2.1.8 include:
1. --build has improved and KING pedigree reconstruction can now incorporate 2nd-degree relatedness inference.
2. Drawing pedigrees is now feasible through --rplot or --build --rplot, if R library "kinship2" is installed.
3. One change of memory management improves performance of multiple analyses for biobank-level data.
Linux-king216.tar.gz For GNU/LINUX 64 bit systems
Ubuntu-king216.tar.gz For Ubuntu 64 bit systems
Mac-king216.tar.gz For Mac Workstations
KING2.1.6code.tar.gz C++ source code
Major changes in KING 2.1.6 include:
1. Much less computer memory is required than in previous versions.
As a result, --lessmem option is retired
2. Max sample size restriction is removed and now
--duplicate and --related inferences are applicable to datasets of any size, e.g.,N>10,000,000
3. Multiple datasets can now be read in together, without worrying about allele strands. e.g.,
king -b ex1,ex2 --duplicate
4. KING now prints out % of progress on the screen while waiting for large data processing
5. --build can now take a rather small sample size (which crashes in previous versions when N<100)
The largest dataset we have successfully analyzed on a single server with both --duplicate and --related options consists of >10,000,000 samples
(i.e., > 50,000,000,000,000 pairs!).
I believe both options work easily (in minutes) on any biobank data nowadays and in the near future.
Linux-king215.tar.gz For GNU/LINUX 64 bit systems
Ubuntu-king215.tar.gz For Ubuntu 64 bit systems
Mac-king215.tar.gz For Mac Workstations (Single-thread version here for Mac without OMP library support)
KING2.1.5code.tar.gz C++ source code
New options are added in KING 2.1.5, including
1. --sexchr is newly added. Many non-human species can now be analyzed in KING. Default sexchr is 23
2. --risk is fixed and substantially improved. Tutorials for genetic risk score (GRS) and risk prediction are now available
3. --ibdseg can now analyze up to 700,000 samples instead of 256,000 samples as in KING 2.1.4
Comparing to the previous version of KING 2.1.4, this update may be not crucial for many users.
Linux-king214.tar.gz For GNU/LINUX 64 bit systems
Ubuntu-king214.tar.gz For Ubuntu 64 bit systems
Mac-king214.tar.gz For Mac Workstations (Single-thread version here for Mac without OMP library support)
KING2.1.4code.tar.gz C++ source code
A number of algorithms are further improved in KING 2.1.4, including --related, --rplot, etc.
KING can now accurately infer up to the 4th-degree relatedness (--related, --ibdseg),
while the original KING method (--kinship) remains accurate up to the second-degree relatedness (even in the presence of admixture).
Linux-king213.tar.gz For GNU/LINUX 64 bit systems
Ubuntu-king213.tar.gz For Ubuntu 64 bit systems
Mac-king213.tar.gz For Mac Workstations (Single-thread version here for Mac without OMP library support)
KING2.1.3code.tar.gz C++ source code
A number of algorithms are improved in KING 2.1.3, including --ibdseg, --related and --roh.
A few bugs have been fixed (including --cluster --bySNP, --build etc).
Linux-king212.tar.gz For GNU/LINUX 64 bit systems
Ubuntu-king212.tar.gz For Ubuntu 64 bit systems
Mac-king212.tar.gz For Mac Workstations
KING2.1.2code.tar.gz C++ source code
The IBD segment algorithm is improved in KING 2.1.2.
Linux-king21.tar.gz For GNU/LINUX 64 bit systems
Ubuntu-king21.tar.gz For Ubuntu 64 bit systems
Mac-king21.tar.gz For Mac Workstations
KING2.1code.tar.gz C++ source code
KING 2.1 is a major update. New features include:
1. Improved multi-core parallel computing algorithms
2. --related option is integrative and more accurate
3. New functions added: --ibdseg for pair-wise IBD segment inference
Linux-king20.tar.gz For GNU/LINUX 64 bit systems
Mac-king20.tar.gz For Mac Workstations
Ubuntu-king20.tar.gz For Ubuntu 32 bit systems
Mac-king20-single-thread.gz For Mac users without OpenMP support
KING2.0code.tar.gz C++ source code
KING 2.0 is a major update. New features include:
1. Multi-core parallel computing
2. Improved algorithms. Many KING analyses can now be done in a few seconds (assuming ~10,000 samples)
3. New functions added, e.g., --tdt, --mtscore
Linux-king19.tar.gz For GNU/LINUX 64 bit systems
A much faster algorithm is implemented in this beta version. It is now practical to identify all close relatives in 1,000,000 individuals.
Please try "king --duplicate", "king --related" or "king --related --degree 2" for inference of duplicates, close relationships up to the first degree, and close relationships up to the second degree, respectively.
Linux-king142.tar.gz For GNU/LINUX 64 bit systems
Due to a popular request, "king --pca --projection" is now available in the released version. This option allows "Hapmap rooted" analysis, in which each study sample is projected to the reference (e.g., Hapmap 3) PC space. Note reference samples need to be indicated as 1 in the phenotype (6th) column, and the study samples need to 2.
Linux-king.tar14.gz For GNU/LINUX 64 bit systems
Windows-king14.zip For Windows Workstations
Mac-king14.tar.gz For Mac Workstations
Ubuntu-king14.tar.gz For Ubuntu Workstations
New implementation: "king --unrelated" or "king --unrelated --degree 2" extracts a list of unrelated individuals
"king --related" or "king --related --degree 2" is now much faster than
before. With this newly implemented algorithm, there shouldn't be any
computational difficulty for KING to
analyze >100,000 individuals.
Linux-king.tar11.gz For GNU/LINUX 64 bit systems
Windows-king11.zip For Windows Workstations
Algorithms for relationship inference and population structure analysis are further optimized.
"--kinship", "--ibs --kinship", "--homo", and "--mds" analyses are all faster than before. One bug in
the "--homo" analysis is fixed.
Linux-king107.tar.gz For GNU/LINUX 64 bit systems
Windows-king107.zip For Windows Workstations
A minor bug is fixed in KING version 1.0.7. This bug did not affect any
of the KING analyses. The bug involved slight change of SNP positions when
new data were generated (through --binary, --plink or --merlin) and when MERLIN
format data were used as input.
Linux-king.tar105.gz For GNU/LINUX 64 bit systems
Windows-king105.zip For Windows Workstations
KING version 1.0.5 is now able to properly perform population structure analysis (MDS) when a PLINK binary dataset is used.
Linux-king.tar104.gz For GNU/LINUX 64 bit systems
Windows-king104.zip For Windows Workstations
One bug in version 1.0.3 has now been fixed in 1.0.4. Versions older than 1.0.3 did not have
this bug. The bug involves reading Merlin format data without the use the map file. KING 1.0.3 in this
case reported that no autosome SNPs were found.
Linux-king.tar103.gz For GNU/LINUX 64 bit systems
Windows-king103.zip For Windows Workstations
Mac-king103.tar.gz For Mac Workstations
Ubuntu-king103.tar.gz For Ubuntu 32 bit systems
One bug is fixed. This bug occurred when 1) MERLIN-format input is used; 2) Some SNPs are on X-chromosome; and
3) --individual or --homo analysis is specified. The previous version of KING was running correctly when only one or two conditions held.
Linux-king.tar102.gz For GNU/LINUX 64 bit systems
Windows-king102.zip For Windows Workstations
Mac-king102.tar.gz For Mac Workstations
Ubuntu-king102.tar.gz For Ubuntu 32 bit systems
Bugs related to format convertion were identified in previous versions. Luckily none of these
affects any of the existing analyses (relationship inference, structure analysis, etc.). If your analysis involves
using "--merlin" (saving the data in MERLIN format) previously, please try the new version again.
Linux-king.tar101.gz For GNU/LINUX 64 bit systems
Windows-king101.zip For Windows Workstations
Mac-king101.tar.gz For Mac Workstations
Ubuntu-king101.tar.gz For Ubuntu 32 bit systems
Linux-king100.tar.gz For GNU/LINUX 64bit systems
Windows-king100.zip For Windows Workstations
Mac-king100.tar.gz For Mac Workstations
Ubuntu-king100.tar.gz For Ubuntu 32 bit systems
If you find KING helpful, please cite
Manichaikul A, Mychaleckyj JC, Rich SS, Daly K, Sale M, Chen WM
(2010) Robust relationship inference in genome wide association studies.
Bioinformatics 26(22):2867-2873 [Abstract]
Last updated: March 23, 2021 by Wei-Min Chen