FastNGSadmix - Revision history

Emil: /* Quick start - 1000 genomes reference panel */

2018-04-15T10:48:53Z

Quick start - 1000 genomes reference panel

← Older revision		Revision as of 12:48, 15 April 2018
Line 116:		Line 116:

	Please be aware if you are not a server with a lot of RAM, that the PCA part might consume a lot of RAM,		Please be aware if you are not a server with a lot of RAM, that the PCA part might consume a lot of RAM,
	~~that is~~ if you are including all ref sites for the PCA (changing the settings to -overlapRef 0).		especially if you are including all ref sites for the PCA (changing the settings to -overlapRef 0).

	=Input Files=		=Input Files=

Emil: /* Quick start - 1000 genomes reference panel */

2018-04-15T10:34:31Z

Quick start - 1000 genomes reference panel

← Older revision		Revision as of 12:34, 15 April 2018
Line 108:		Line 108:

	#Estimate admixture proportions from the $GL genotype likelihoods file		#Estimate admixture proportions from the $GL genotype likelihoods file
	./fastNGSadmix -likes CEU.low_coverage.beagle -fname $REF -Nname $NIND -out CEU.low_coverage -whichPops ~~all~~		./fastNGSadmix -likes CEU.low_coverage.beagle.gz -fname $REF -Nname $NIND -out CEU.low_coverage -whichPops CEU,CHB,YRI,PEL

	#do PCA based on the calculated admixture proportions from the previous command, using the $GL genotype likelihoods file		#do PCA based on the calculated admixture proportions from the previous command, using the $GL genotype likelihoods file
	Rscript R/fastNGSadmixPCA.R -likes CEU.low_coverage.beagle -qopt CEU.low_coverage.qopt -out CEU.low_coverage -ref $GENO		Rscript R/fastNGSadmixPCA.R -likes CEU.low_coverage.beagle.gz -qopt CEU.low_coverage.qopt -out CEU.low_coverage -ref $GENO

	</pre>		</pre>

Emil at 09:50, 15 April 2018

2018-04-15T09:50:20Z

Emil: /* Quick start - 1000 genomes reference panel */

2018-04-15T09:45:09Z

Quick start - 1000 genomes reference panel

← Older revision		Revision as of 11:45, 15 April 2018
Line 116:		Line 116:

	Please be aware if you are not a server with a lot of RAM, that the PCA part might consume a lot of RAM,		Please be aware if you are not a server with a lot of RAM, that the PCA part might consume a lot of RAM,
	that is if you are including all ref sites for the PCA (~~chang~~ the ~~setting~~ to -overlapRef 0).		that is if you are including all ref sites for the PCA (changing the settings to -overlapRef 0).

	=Input Files=		=Input Files=

Emil: /* Quick start - 1000 genomes reference panel */

2018-04-15T09:44:33Z

Quick start - 1000 genomes reference panel

← Older revision		Revision as of 11:44, 15 April 2018
Line 110:		Line 110:
	./fastNGSadmix -likes CEU.low_coverage.beagle -fname $REF -Nname $NIND -out CEU.low_coverage -whichPops all		./fastNGSadmix -likes CEU.low_coverage.beagle -fname $REF -Nname $NIND -out CEU.low_coverage -whichPops all

	#do PCA based on the calculated admixture proportions from the previous command, using the $GL genotype likelihoods		#do PCA based on the calculated admixture proportions from the previous command, using the $GL genotype likelihoods file
	Rscript R/fastNGSadmixPCA.R -likes CEU.low_coverage.beagle -qopt CEU.low_coverage.qopt -out CEU.low_coverage -ref $GENO		Rscript R/fastNGSadmixPCA.R -likes CEU.low_coverage.beagle -qopt CEU.low_coverage.qopt -out CEU.low_coverage -ref $GENO

Emil at 09:44, 15 April 2018

2018-04-15T09:44:14Z

← Older revision		Revision as of 11:44, 15 April 2018
Line 110:		Line 110:
	./fastNGSadmix -likes CEU.low_coverage.beagle -fname $REF -Nname $NIND -out CEU.low_coverage -whichPops all		./fastNGSadmix -likes CEU.low_coverage.beagle -fname $REF -Nname $NIND -out CEU.low_coverage -whichPops all

	#~~perform~~ PCA based on the calculated admixture proportions from the previous command ~~and~~ using the $GL genotype likelihoods ~~file~~		#do PCA based on the calculated admixture proportions from the previous command, using the $GL genotype likelihoods
	Rscript R/fastNGSadmixPCA.R -likes CEU.low_coverage.beagle -qopt CEU.low_coverage.qopt -out CEU.low_coverage -ref $GENO		Rscript R/fastNGSadmixPCA.R -likes CEU.low_coverage.beagle -qopt CEU.low_coverage.qopt -out CEU.low_coverage -ref $GENO

Emil: /* Quick start - 1000 genomes reference panel */

2018-04-15T09:42:53Z

Quick start - 1000 genomes reference panel

← Older revision		Revision as of 11:42, 15 April 2018
Line 68:		Line 68:
	Utah Residents (CEPH) with Northern and Western European Ancestry (CEU), Han Chinese (CHB), Finnish (FIN), Peruvian (PEL), Punjabi (PJL) and Yoruba (YRI), in order to cover most of the worldwide genetic diversity.		Utah Residents (CEPH) with Northern and Western European Ancestry (CEU), Han Chinese (CHB), Finnish (FIN), Peruvian (PEL), Punjabi (PJL) and Yoruba (YRI), in order to cover most of the worldwide genetic diversity.
	For the Peruvians I made sure to only select individauls with no European admixture.		For the Peruvians I made sure to only select individauls with no European admixture.
			The sites were filtered so MAF > 0.05 and missingness < 0.01 and only SNPs were included, leaving 6 676 750 sites.

	First download these files:		First download these files:
Line 93:		Line 94:
	#frequencies from reference panel		#frequencies from reference panel
	REF=data1000genomes/refPanel_1000genomesRefPanel.txt		REF=data1000genomes/refPanel_1000genomesRefPanel.txt

			#sites file to be used with ANGSD
			SITES=data1000genomes/1000genomesRefPanel.sites

	#number of individuals in each population of reference panel		#number of individuals in each population of reference panel
	NIND=data1000genomes/nInd_1000genomesRefPanel.txt		NIND=data1000genomes/nInd_1000genomesRefPanel.txt

			#genotypes of the 1000 genomes data - for the PCA part
			GENO=data1000genomes/1000genomesRefPanel

	#use ANGSD for generating a beagle file with the genotypes likelihoods		#use ANGSD for generating a beagle file with the genotypes likelihoods
	./angsd		./angsd -i $BAM -GL 2 -sites $SITES -doGlf 2 -doMajorMinor 3 -minMapQ 30 -minQ 20 -doDepth 1 -doCounts 1 -out CEU.low_coverage

	#Estimate admixture proportions from the $GL genotype likelihoods file		#Estimate admixture proportions from the $GL genotype likelihoods file
	./fastNGSadmix -likes ~~$GL~~ -fname $REF -Nname $NIND -out ~~yriFrenchHan_depth05~~ -whichPops ~~French,Han,Yoruba~~		./fastNGSadmix -likes CEU.low_coverage.beagle -fname $REF -Nname $NIND -out CEU.low_coverage -whichPops all

	#perform PCA based on the calculated admixture proportions from the previous command and using the $GL genotype likelihoods file		#perform PCA based on the calculated admixture proportions from the previous command and using the $GL genotype likelihoods file
	Rscript R/fastNGSadmixPCA.R -likes ~~$GL~~ -qopt ~~yriFrenchHan_depth05~~.qopt -out ~~yriFrenchHan_depth05~~ -ref ~~data/humanOrigins_7worldPops~~		Rscript R/fastNGSadmixPCA.R -likes CEU.low_coverage.beagle -qopt CEU.low_coverage.qopt -out CEU.low_coverage -ref $GENO

	</pre>		</pre>

			Please be aware if you are not a server with a lot of RAM, that the PCA part might consume a lot of RAM,
			that is if you are including all ref sites for the PCA (chang the setting to -overlapRef 0).

	=Input Files=		=Input Files=

Emil at 09:33, 15 April 2018

2018-04-15T09:33:58Z

← Older revision		Revision as of 11:33, 15 April 2018
Line 61:		Line 61:

	</pre>		</pre>


			==Quick start - 1000 genomes reference panel==

			A reference panel was created using the 1000 genomes data, taking 20 individuls from the following populations:
			Utah Residents (CEPH) with Northern and Western European Ancestry (CEU), Han Chinese (CHB), Finnish (FIN), Peruvian (PEL), Punjabi (PJL) and Yoruba (YRI), in order to cover most of the worldwide genetic diversity.
			For the Peruvians I made sure to only select individauls with no European admixture.

			First download these files:

			<pre>
			wget popgen.dk/software/download/fastNGSadmix/data1000genomes.tar.gz
			wget popgen.dk/software/download/fastNGSadmix/example.tar.gz
			</pre>

			Use curl if you are on a MAC

			They can be unpacked thus:

			<pre>
			tar -xzf data1000genomes.tar.gz
			tar -xzf example.tar.gz
			</pre>

			And then you can easily run your analysis on a .bam file - requires [http://popgen.dk/angsd ANGSD]:

			<pre>
			#example bam file with genotype data
			BAM=example/smallNA12874.mapped.ILLUMINA.bwa.CEU.low_coverage.20130415.bam

			#frequencies from reference panel
			REF=data1000genomes/refPanel_1000genomesRefPanel.txt

			#number of individuals in each population of reference panel
			NIND=data1000genomes/nInd_1000genomesRefPanel.txt

			#use ANGSD for generating a beagle file with the genotypes likelihoods
			./angsd

			#Estimate admixture proportions from the $GL genotype likelihoods file
			./fastNGSadmix -likes $GL -fname $REF -Nname $NIND -out yriFrenchHan_depth05 -whichPops French,Han,Yoruba

			#perform PCA based on the calculated admixture proportions from the previous command and using the $GL genotype likelihoods file
			Rscript R/fastNGSadmixPCA.R -likes $GL -qopt yriFrenchHan_depth05.qopt -out yriFrenchHan_depth05 -ref data/humanOrigins_7worldPops

			</pre>


	=Input Files=		=Input Files=

Emil: /* Input Files */

2018-02-03T14:54:09Z

Input Files

← Older revision		Revision as of 16:54, 3 February 2018
Line 90:		Line 90:

	Where the marker column is '''recommended to be chr_pos''', instead of rs ID - but custom ids can be used.		Where the marker column is '''recommended to be chr_pos''', instead of rs ID - but custom ids can be used.

	If one wants to analyse a haploid organism, one should create a regular beagle file (with 6 columns),		If one wants to analyse a haploid organism, one should create a regular beagle file (with 6 columns),
	but only fill out the two first genotype likelihood columns, for respectively specifying the genotype likelihood of allele1 and allele2.		but only fill out the two first genotype likelihood columns, for respectively specifying the genotype likelihood of allele1 and allele2.

			Example of haploid data in a beagle file:

			<pre>
			marker allele1 allele2 Ind0 Ind0 Ind0
			1_752566 A G 0.50 0.50 0
			1_842013 T G 0.99 0.01 0
			1_891021 A G 0.99 0.01 0
			...
			</pre>

	The reference panel looks like this:		The reference panel looks like this:

Emil at 14:50, 3 February 2018

2018-02-03T14:50:17Z

← Older revision		Revision as of 16:50, 3 February 2018
Line 90:		Line 90:

	Where the marker column is '''recommended to be chr_pos''', instead of rs ID - but custom ids can be used.		Where the marker column is '''recommended to be chr_pos''', instead of rs ID - but custom ids can be used.
			If one wants to analyse a haploid organism, one should create a regular beagle file (with 6 columns),
			but only fill out the two first genotype likelihood columns, for respectively specifying the genotype likelihood of allele1 and allele2.

	The reference panel looks like this:		The reference panel looks like this:
Line 194:		Line 196:

	'''Optional:'''		'''Optional:'''
	; -haploid [FLAG] Raise this flag if haploid organism being analyzed, first two cols of beagle file will be used - does not work for plink! Only write "-haploid".		; -haploid [FLAG] (Only write '-haploid' in command line)
			Raise this flag if haploid organism being analyzed, first two cols of beagle file will be used - does not work for plink! Only write "-haploid".
	; -out [char*]		; -out [char*]
	Prefix for the output files .qopt and .log.		Prefix for the output files .qopt and .log.
Line 648:		Line 651:
	=Log=		=Log=
	version 1.0: Stable initially released version of fastNGSadmix - 07-09-2017<br>		version 1.0: Stable initially released version of fastNGSadmix - 07-09-2017<br>
	version 1.1: Beagle input and reference panel do not have to be sorted genomically, only find overlap based on ID - 23-09-2017		version 1.1: Beagle input and reference panel do not have to be sorted genomically, only find overlap based on ID - 23-09-2017<br>
			version 1.2: Can also handle and analyse haploid data, only with beagle input files - 03-02-2018

← Older revision		Revision as of 11:50, 15 April 2018
Line 49:		Line 49:

	#Estimate admixture proportions from the $GL genotype likelihoods file		#Estimate admixture proportions from the $GL genotype likelihoods file
	./fastNGSadmix -likes $GL -fname $REF -Nname $NIND -out yriFrenchHan_depth05 -whichPops French,Han,Yoruba		./fastNGSadmix -likes $GL -fname $REF -Nname $NIND -out yriFrenchHan_depth05 -whichPops French,Han,Yoruba

	#perform PCA based on the calculated admixture proportions from the previous command and using the $GL genotype likelihoods file		#perform PCA based on the calculated admixture proportions from the previous command and using the $GL genotype likelihoods file
	Rscript R/fastNGSadmixPCA.R -likes $GL -qopt yriFrenchHan_depth05.qopt -out yriFrenchHan_depth05 -ref data/humanOrigins_7worldPops		Rscript R/fastNGSadmixPCA.R -likes $GL -qopt yriFrenchHan_depth05.qopt -out yriFrenchHan_depth05 -ref data/humanOrigins_7worldPops

	#Estimate admixture proportions from the $PLINKFILE plink file		#Estimate admixture proportions from the $PLINKFILE plink file
	./fastNGSadmix -plink $PLINKFILE -fname $REF -Nname $NIND -out NA20502_TSI -whichPops French,Han,Yoruba		./fastNGSadmix -plink $PLINKFILE -fname $REF -Nname $NIND -out NA20502_TSI -whichPops French,Han,Yoruba

	#perform PCA based on the calculated admixture proportions from the previous command and using the $PLINKFILE plink file		#perform PCA based on the calculated admixture proportions from the previous command and using the $PLINKFILE plink file
	Rscript R/fastNGSadmixPCA.R -plinkFile $PLINKFILE -qopt NA20502_TSI.qopt -out NA20502_TSI -ref data/humanOrigins_7worldPops		Rscript R/fastNGSadmixPCA.R -plinkFile $PLINKFILE -qopt NA20502_TSI.qopt -out NA20502_TSI -ref data/humanOrigins_7worldPops

	</pre>		</pre>
Line 439:		Line 439:
	The method can also be run only using the estimated prior, that means calculating the genotype probability given only the admixture proportions and population specific frequencies		The method can also be run only using the estimated prior, that means calculating the genotype probability given only the admixture proportions and population specific frequencies
	and NOT the sequencing data, for more see supplementaries of Jørsboe et al., 2017. This is done by setting '-onlyPrior 1'.		and NOT the sequencing data, for more see supplementaries of Jørsboe et al., 2017. This is done by setting '-onlyPrior 1'.
	Setting '-overlapRef 1' means that the covariance matrix for the reference panel genotypes is calculated ~~only on overlapping~~ markers ~~with input~~ (per default ~~all~~ markers are used for calculating the reference panel covariance matrix).		Setting '-overlapRef 0' means that the covariance matrix for the reference panel genotypes is calculated using all markers in the reference genotypes (per default only overlapping markers are used for calculating the reference panel covariance matrix).