Back to index.html.

Files used in genomic analysis

In this section, we review the files required in the genomic analysis by the BLUPF90 programs. The programs read and check these files in the preparation step as well as the rest of genomic-analysis steps.

The BLUPF90 programs need at least two files for genomic data: an SNP marker file and a cross-reference file. In this section, we describe the formats of these two files. We also present some other files that may be needed for marker quality control.

SNP file

The BLUPF90 programs can accept two kinds of SNP files. One contains only integer genotypes (for example, from SNP chips), and the other contains real numbers as gene content (possibly from imputation software, genotyping-by-sequencing, or low-depth sequencing). First, we consider the former; we call it an SNP-marker file.

SNP marker file

An SNP-marker file is a text file containing two fields: the animal ID (possibly alphanumeric) in the first field and its genotypes in the second field. We show the first and last five lines from an example data set presented in the previous chapter (limited to the first 30 characters of each line to save space).

 8003   211011112112012000211002
 8007   212011111111012011111012
 8016   200120111021012111102121
 8019   112211021121202111011210
 8020   211110101120002021002122
   (skip)
13496   101001212012010210212201
13497   101112111021020120222111
13498   200000220202202000022222
13499   011011111111020111121112
13500   101020112122021000221001

The format was already described in the previous chapter.

Gene content file

The BLUPF90 programs accept another format. Here, we call it a gene content file. A gene content file may have any number of fields, subject to the following rules.

• The animal ID must be in the first field, and the gene content at each locus must be in the second or a later field.
• Adjacent gene-content values can be separated by whitespace, or no spaces may be used.
• No headers, no comments, no other fields are allowed. The data must start on the first line.
• In animal IDs, letters, numbers, and symbols (possibly ASCII) are acceptable. The link to the renumbered pedigree is provided in the cross-reference file below.
• Gene content should be an integer, a floating-point value (for example, 3.14), or an exponential expression (for example, 0.314E+01) represented as a real number. All markers must use the same format.
• Missing gene content is not allowed, and all animals must have the same number of gene-content values. Missing gene content should be imputed before the analyses.
• The minimum number of markers is 50. Fewer markers are not acceptable.

Cross-reference (XrefID) file

The BLUPF90 programs need a cross-reference file. This file is automatically generated by RENUMF90. It relates a renumbered ID to the original ID for genotyped animals. Again, we show the first five and last five lines of the actual cross-reference file presented in the previous chapter.

6127 8003
13570 8007
406 8016
10802 8019
10924 8020
  (skip)
8585 13496
8941 13497
9369 13498
9753 13499
9905 13500

This file simply contains two fields: the first is the renumbered ID (the same as in the pedigree file), and the second is the original ID (the same as in the marker file). The two fields can be separated by at least one space. A tab is not allowed. The order of the animals must be the same as in the SNP file. Again, this file is generated by RENUMF90, and the user should not edit it unless there is a reason to do so.

Allele frequency file (optional)

An allele frequency file contains the allele frequency for each marker. This file is optional because allele frequencies are calculated from the current SNP file by default. If you want to use external information, you can provide a file containing the allele frequencies. Different allele frequencies may change the characteristics of \(\mathbf{G}\).

Here is an example file. It contains allele frequencies only for the first five markers.

1     0.711667
2     0.328000
3     0.422000
4     0.157000
5     0.492333

This file contains two columns:

1. The position of the marker (from 1 to the maximum number of markers)
2. Allele frequency as a real value (ranging from 0 to 1).

Note that the genomic preparation step usually creates this file containing the allele frequencies calculated from the current SNP file (the file name is freqdata.count).

Map file (optional)

The map file relates a marker to a chromosome, a physical location, and a specific name. This file is needed only when you perform comprehensive quality control and GWAS with ssGBLUP. It should contain at least three fields separated by at least one whitespace character. The fourth column is optional and contains the marker name. Here we show an example that includes the fourth field. In addition, the first line must be a header with three space-separated items: SNP_ID, CHR, and POS.

The following example shows the first 4 lines only.

SNP_ID  CHR     POS       SNP_NAME
     1    1     127     SNP-CODE-1
     2    1     652     SNP-CODE-2
     3    1    1022     SNP-CODE-3

This file follows the rules.

• The first line must have a header with three space-separated columns: SNP_ID, CHR, and POS.
• The first three fields should contain integer values: the first is a marker number, the second specifies the chromosome number, and the third represents the physical location on the chromosome. The marker number is just an integer, not necessarily consecutive, used for external data manipulation; it is not actually used by the program.
• The sex chromosome (X) can be present, but it must also be represented by an integer. The code for the X chromosome is 0 by default.
• The optional fourth column can contain any letters, numbers, and symbols (possibly ASCII), up to 50 characters.

Weight for SNP (optional)

The preparation step can create an alternative \(\mathbf{G}\) by weighting each SNP marker. This weighted matrix is especially useful for GWAS or related analyses with POSTGSF90. If you supply this file, the weighted \(\mathbf{G}\) will be calculated. Otherwise, all weights are set to 1, meaning that no specific weight is assigned to any SNP marker. This file is optional.

The file contains only one column of real values. Each row corresponds to one SNP marker: the first line contains the weight for marker 1, the second line contains the weight for marker 2, and so on. The number of lines in this file should not exceed the number of markers.

Back to index.html.