Colocalization results format

The colocalization data contains data that was acquired by running a colocalization analysis between FinnGen data and other datasets.

The colocalization results are tab-separated files. The results are separated to different files, depending on the type of data produced.

More information about data sources, phenotypes and loci can be found in the Colocalization in FinnGen page.

For information about the results format for colocalization results before DF13, using the older colocalization pipeline, see Colocalization results in colocalization before DF13.

The files that are created from the susie-coloc pipeline are:

File	Description
colocQC.tsv.gz	Colocalization summaries between FinnGen endpoints and other resources
coloc.credsets.tsv.gz	Credible set variants that were involved in filtered colocalization results
coloc.H4_tables.tsv.gz	Per-variant H4 posterior probabilities for variants in filtered colocalization results
unfiltered_summaries/FinnGen-R12-GWAS-----{Other resource}.sum.unfiltered.tsv.gz	Unfiltered colocalization summaries for a single resource

Colocalization summaries

The colocalization summaries contain one colocalization per row. The colocalizations have been filtered to not include all of the possible colocalization pairs. The files contain the following columns:

Field	Description
dataset1	name of dataset 1
dataset2	name of dataset 2
trait1	name of trait 1
trait2	name of trait 2
region1	finemapped region of 1st trait
region2	finemapped region of 2nd trait
cs1	credible set index of 1st trait
cs2	credible set index of 2nd trait
nsnps	Amount of snps in overlap of regions 1 and 2
hit1	variant that coloc predicted to be the most likely causal variant in trait 1
hit2	variant that coloc predicted to be the most likely causal variant in trait 2
PP.H0.abf	posterior probability of hypothesis 0: No genetic association in either trait
PP.H1.abf	posterior probability of hypothesis 1: Genetic association in trait 1 only
PP.H2.abf	posterior probability of hypothesis 2: Genetic association in trait 2 only
PP.H3.abf	posterior probability of hypothesis 3: Both traits associated, but with different causal variants
PP.H4.abf	posterior probability of hypothesis 4: Both traits associated and share single causal variant
low_purity1	If credible set 1 has low purity. In short, not all variants in signal have high LD. See finemapping documentation for more information
low_purity2	If credible set 2 has low purity. In short, not all variants in signal have high LD. See finemapping documentation for more information
nsnps1	Amount of snps in region 1
nsnps2	Amount of snps in region 2
cs1_log10bf	log10 bayes factor of credible set 1
cs2_log10bf	log10 bayes factor of credible set 2
clpp	CLPP between credible sets
clpa	CLPA between credible sets
cs1_size	Number of variants in credible set 1
cs2_size	Number of variants in credible set 2
cs_overlap	Overlapping credible set variants in both credible sets
topInOverlap	Whether the maximum PIP variant was in overlap of regions or not, for both traits
probmass_1	Amount of PIP mass of credible set 1 that was in region that was overlapping between the regions. Low value implies that the colocalization can not capture the signal in trait 1.
probmass_2	Amount of PIP mass of credible set 2 that was in region that was overlapping between the regions. Low value implies that the colocalization can not capture the signal in trait 2.
hit1_info	Information about the most likely causal variant as predicted by coloc. Values are beta and p-value separated by comma
hit2_info	Information about the most likely causal variant as predicted by coloc. Values are beta and p-value separated by comma
colocRes	Name of the colocalization intermediate file that this row is in.

credible set variants in colocalizations

The credible set variants file contains the credible set variants for all of the credible sets that appear in the filtered colocalizations. They are listed only once per credible set. The file contains the following columns:

Field	Description
trait	Trait the credible set variant is from
region	Finemapping region
rsid	variant identifier, in format chromosome_position_reference allele_alternate allele
cs	credible set index
low_purity	Whether variants in credible set were not in high LD.
p	p-value of variant association
beta	effect size of variant association
se	standard error of variant association effect
cs_specific_prob	posterior inclusion probability of this variant in credible set
dataset	Dataset identifier

Posterior probability of colocalization table

The H4 table contains the posterior probability of colocalization for all of the variants that were included in the overlap of regions of the colocalization. These variants are only available for the filtered colocalizations int he colocalization summary.

Field	Description
dataset1	dataset of trait1
dataset2	dataset of trait2
trait1	trait of first credible set
trait2	trait of second credible set
region1	finemapped region of first trait
region2	finemapped region of second trait
cs1	credible set index for credible set 1
cs2	credible set index for credible set 2
snp	snp identifier, in format chromosome_position_reference allele_alternate allele
SNP.PP.H4	Posterior probability of this variant being the causal variant in this colocalization, given that hypothesis 4 is true

Unfiltered colocalization summaries

These files contain the colocalization summaries that are not filtered. The columns are largely the same as in the filtered colocalization summaries, except that the dataset columns are not present, as those can be taken from the filename, and the colocRes column is missing as well, since it is the filename.The results contains the following columns:

Field	Description
trait1	name of trait 1
trait2	name of trait 2
region1	finemapped region of 1st trait
region2	finemapped region of 2nd trait
cs1	credible set index of 1st trait
cs2	credible set index of 2nd trait
nsnps	Amount of snps in overlap of regions 1 and 2
hit1	variant that coloc predicted to be the most likely causal variant in trait 1
hit2	variant that coloc predicted to be the most likely causal variant in trait 2
PP.H0.abf	posterior probability of hypothesis 0: No genetic association in either trait
PP.H1.abf	posterior probability of hypothesis 1: Genetic association in trait 1 only
PP.H2.abf	posterior probability of hypothesis 2: Genetic association in trait 2 only
PP.H3.abf	posterior probability of hypothesis 3: Both traits associated, but with different causal variants
PP.H4.abf	posterior probability of hypothesis 4: Both traits associated and share single causal variant
low_purity1	If credible set 1 has low purity. In short, not all variants in signal have high LD. See finemapping documentation for more information
low_purity2	If credible set 2 has low purity. In short, not all variants in signal have high LD. See finemapping documentation for more information
nsnps1	Amount of snps in region 1
nsnps2	Amount of snps in region 2
cs1_log10bf	log10 bayes factor of credible set 1
cs2_log10bf	log10 bayes factor of credible set 2
clpp	CLPP between credible sets
clpa	CLPA between credible sets
cs1_size	Number of variants in credible set 1
cs2_size	Number of variants in credible set 2
cs_overlap	Overlapping credible set variants in both credible sets
topInOverlap	Whether the maximum PIP variant was in overlap of regions or not, for both traits
probmass_1	Amount of PIP mass of credible set 1 that was in region that was overlapping between the regions. Low value implies that the colocalization can not capture the signal in trait 1.
probmass_2	Amount of PIP mass of credible set 2 that was in region that was overlapping between the regions. Low value implies that the colocalization can not capture the signal in trait 2.
hit1_info	Information about the most likely causal variant as predicted by coloc. Values are beta and p-value separated by comma
hit2_info	Information about the most likely causal variant as predicted by coloc. Values are beta and p-value separated by comma

More information about the methods can be found in th Colocalization in FinnGen page, as well as the methods document:

/finngen/library-green/finngen_R12_analysis_data/colocalization/methods.pdf

More documentation about the data, including the columns and their descriptions, can be found in the release notes.

Read more about Colocalization and Colocalization in FinnGen