LAAA, the pipeline for Local Ancestry and Allelic Association, consists of the following four stages:

1. Format conversion
2. Local ancestry inference
3. Post-inference processing
4. Two-step testing

Note: Examples used in this tutorial as well as a more detailed README can be found in the downloadable package.

STAGE 1

Convert sample genotypes in plink format and reference haplotypes to HAPMIX required input format and generate parameter file.
The following input files are required at stage 1:

• A ped and a map file containing genotypes from study sample in plink format.
• A haps file containing phased reference haplotypes (e.g., HapMapII or HapMapIII phased haplotype files, or 1000 Genomes phased haplotype files)
• A snps file containing all markers present in the above hap file (e.g., the snps files come with HapMapII phased haplotype files)
• A rate file consisting of the following 5 columns (e.g., the recombination rate files come with HapMapII phased haplotype files)
  • 1st column: marker ID
  • 2nd column: chromosome
  • 3rd column: physical position
  • 4th column: recombination rate (cM/Mb)
  • 5th column: genetic distance (cM)

Note:

• Genetic distance is required for input genotypes.
• All markers in the reference files must be included in the rate file.
• All markers in teh genotype files must be inclded in the reference and rate files.

The input files are specified using the following command line options:

• chromosome number: -chr
• ped file: -ped
• map file: -map
• haps file from population 1: -hap1
• snps file from population 1: -snps1
• haps file from population 2: -hap2
• snps file from population 2: -snps2
• rate file: -rate
• output directory: -outdir

Example command line (details on this example can be found in the README included in the downloadable package):
./convert_format.pl \
-chr 22 \
-ped example.ped \
-map example.map \
-hap1 CEU.hap \
-snps1 CEU.snps \
-hap2 YRI.hap \
-snps2 YRI.snps \
-rate rate.txt \
-outdir out_hapmix

Output:
Input files required by HAPMIX as well as a parameter file. A folder called “RUN” is created for holding results from HAPMIX. For example,

• out_hapmix/AAgenofile.22
• out_hapmix/AAind.22
• out_hapmix/AAsnpfile.22
• out_hapmix/pop1genofile.22
• out_hapmix/pop1snpfile.22
• out_hapmix/pop2genofile.22
• out_hapmix/pop2snpfile.22
• out_hapmix/rates.22
• out_hapmix/hapmix.par.22
• out_hapmix/RUN

STAGE 2

Stage 2 performs local ancestry inference. Output from stage 1 serves as input here. The user can specify to use task scheduler or not. Default is not. Inference is performed on each individual one by one.

Example command line:
./bin/runHapmix.pl out_hapmix/hapmix.par.22

If you are using a task scheduler, three sub-steps are needed:

1. File formatting:
   Example command line:
   ./bin/runHapmix_1_GenerateFiles.pl -parfile out_hapmix/hapmix.par.22
2. Local ancestry inference using HAPMIX; each individual local ancestry inference is submitted as a job. In this example, we use 'bsub' as our task scheduler:
   Example command line:
   ./bin/runHapmix_2_admix2.sh AA 22 out_hapmix bsub
3. Finalizing:
   Example command line:
   ./bin/runHapmix_3_SummarizeFiles.pl -parfile out_hapmix/hapmix.par.22

Output is generated in out_hapmix/RUN as AA.DIPLOID.*.22

STAGE 3

Stage 3 performs post-inference processing to prepare input for the two-step testing procedure.

Input:

• path to HAPMIX output
• chromosome number
• sample size
• output path

Example command line:
# Assume sample size is 20 and output directory is out_testing
Rscript reformat.hapmix.output.R out_hapmix 22 20 out_testing

Output are three files containing matrix of number of minor alleles, number of ancestry-specific alleles and number of ancestry-specific minor alleles. For example:

• AA.geno.22
• AFR.allele.22
• AFR.minor.22

STAGE 4

Perform association analysis using two step testing procedure. Additional input of phenotype and covariates files are required.

Input:

• path to the three matrix generated in step 3
• chromosome number
• traits: two columns; 1st column IID; 2nd column value of trait
• covariates: 1st column: IID; the following columns: value of covariates
• path to association output

Example command line:
Rscript association_2wayAdmixed.R out_testing 22 traits.txt covariates.txt out_testing

The output file is the association results based on two step testing procedure, containing the following 14 columns:

1. SNP ID
2. P value from jointly testing allele effect, ancestry effect and effect heterogeneity
3. Effect size of allele effect
4. Standard error of allele effect
5. Test statistics of allele effect
6. P value of allele effect
7. Effect size of ancestry effect
8. Standard error of ancestry effect
9. Test statistics of ancestry effect
10. P value of ancestry effect
11. Effect size of ancestry specific minor allele
12. Standard error of ancestry specific minor allele
13. Test statistics of ancestry specific minor allele
14. P value of ancestry specific minor allele
15. sample size