Skip to content

Latest commit

 

History

History
143 lines (125 loc) · 3.92 KB

06_Merging_profiles.md

File metadata and controls

143 lines (125 loc) · 3.92 KB
Raw

Merging single-sample profiles into multisample profiles

Once STR profiles have been computed for each sample in the dataset, they must be merged together into a multisample STR profile. Multisample STR profiles are produced by ExpansionHunterDenovo merge command which has the following parameters.

Required parameter Description
--reference The FASTA file to which the reads were aligned
--manifest TSV file with describing each sample
--output-prefix Common prefix for the output files
Optional parameter Description Default
--min-unit-len Shortest repeat unit to consider 2
--max-unit-len Longest repeat unit to consider 20

Manifest files

The manifest file is a tab-delimited file whose columns contain sample id, case/control status, and absolute path to the EH Denovo output file. For example,

sample1	case	/path/to/sample1.str_profile.json
sample2	case	/path/to/sample2.str_profile.json
sample3	case	/path/to/sample3.str_profile.json
sample4	control	/path/to/sample4.str_profile.json
sample5	control	/path/to/sample5.str_profile.json
sample6	control	/path/to/sample6.str_profile.json

Multisample STR profiles

A multisample STR profile is a result of merging multiple single-sample STR profiles. It is a JSON object containing counts of in-repeat reads across all samples for each repeat motif. When two samples have overlapping or nearby regions with anchored in-repeat reads corresponding to the same repeat motif, these regions are merged together. By default, merge operation merges regions located within 500bp of one another.

Example

Let's assume we have two (single-sample) STR profiles that we want to combine. The first profile is the file sample1.str_profile.json containing

{
    "CCG": {
        "AnchoredIrrCount": 7,
        "IrrPairCount": 3,
        "RegionsWithIrrAnchors": {
            "chr1:100-1000": 7
        },
        "RepeatUnit": "CCG",
        "ReadLength": 150,
        "Depth": 30.0
    }
}

And second profile is the file sample2.str_profile.json containing

{
    "AGG": {
        "AnchoredIrrCount": 7,
        "IrrPairCount": 2,
        "RegionsWithIrrAnchors": {
            "chr2:5000-5300": 7
        },
        "RepeatUnit": "AGG"
    },
    "CCG": {
        "AnchoredIrrCount": 3,
        "IrrPairCount": 10,
        "RegionsWithIrrAnchors": {
            "chr1:1150-1200": 3
        },
        "RepeatUnit": "CCG",
        "ReadLength": 150,
        "Depth": 40.0
    }
}

To merge these profiles together we create a manifest file manifest.txt

sample1	case	/path/to/sample1.str_profile.json
sample2	control	/path/to/sample2.str_profile.json

and then run

ExpansionHunterDenovo merge \
  --reference reference.fa \
  --manifest manifest.txt \
  --output-prefix example_dataset

The program generates a file example_dataset.multisample_str_profile.json with the following content

{
    "Counts": {
        "AGG": {
            "IrrPairCounts": {
                "sample2": 2
            },
            "RegionsWithIrrAnchors": {
                "chr2:5000-5300": {
                    "sample2": 7
                }
            }
        },
        "CCG": {
            "IrrPairCounts": {
                "sample1": 3,
                "sample2": 10
            },
            "RegionsWithIrrAnchors": {
                "chr1:100-1200": {
                    "sample1": 7,
                    "sample2": 3
                },
            }
        }
    },
    "Parameters": {
        "Depths": {
            "sample1": 30.0,
            "sample2": 40.0
        },
        "ReadLengths": {
            "sample1": 150,
            "sample2": 150
        }
    }
}