1. Split/align

The split script

First, we will split our .fastq files using the run_BestRadSplit_PstI.sh bash script provided by the Miller Lab.

Note: The script is specific to the PstI cutter, so use SbfI if that is the cutter, etc.

The script gets fed the R1 and R3 .fastq files and we give our output a name, typically the standard name. (e.g. SOMM185_ACTTGA below)

sbatch -p high -t 24:00:00 ../run_BestRadSplit_PstI.sh SOMM185_R1_ACTTGA.fastq SOMM185_R3_ACTTGA.fastq SOMM185_ACTTGA

In a new screen, compress the output of the split script:

srun -t 24:00:00 gzip *ACTTGA.fastq

The align script

Next, we will make two lists of all of the .fastq file names and then paste them together. We’ll then feed this list to a run_align script with a reference genome (available online).

ls *RA* | sed "s/\.fastq\.gz//g" > listA ls *RB* | sed "s/\.fastq\.gz//g" > listB

Here, selected all of the files with RA and RB located inside the file name (ls *RA*) and then substituted (s) .fastq and .gz for a blank space (//). This gives us just the file names without the extensions.

paste listA listB > list_ACTTGA sbatch -t 24:00:00 ../run_align.sh list_ACTTGA ../../Mola/B_imp_ref_genome/GCF_000188095.1_BIMP_2.0_genomic.fna

To align, we combine the two lists into one list (in this case for plate ACTTGA). We then sbatch the run_align.sh script providing it with our list and a reference genome (in this case Bombus impatiens).

A little bit of tidying

To finish the split/align section, I like to organize the output files. This is especially handy when we’re working plate-by-plate rather than in a full run of all plates at once.

mkdir aligned_files filtered_bams slurm_outputs zipped_fastq

Then move the various file types into their respective folders (with mv).

It’s also good to check the length of the bams file (should be 96) using ls filtered_bams/ | wc.

In the end, 4 subdirectories, the list, and the two original .fastq files should remain in the plate directory.

2. Filter to read quality

There are various ways to filter read quality, and some of them will be done later on at the loci level when we call SNPs. However, we can subset to individuals on a minimum number of aligned reads. This helps filter out poorly sequenced individuals who might mess up downstream loci selection and be unusable anyway.

Use the subsample script

need to insert info on subsample script

3. Subset to our sample of interest

For sibship analysis, we want loci that are the most informative within our specific “population” or sample of interest. For example, we know that workers in 2012 won’t be full-sibs of workers in 2016, so we don’t want to SNP call with all of those individuals mixed together. You can likely skip this step for other types of analyses and call SNPs first, then subset for downstream analyses.

Copy over the bamclst_lists.R file (This handy script was originally written by Ryan Peek).

cp ~dir/bamclst_lists.R /dir/.

Refer to the script for exact details.

Input

  • A bamlist, the script assumes you’re using a subsampled list, but this is not necessary.
  • Metadata containing individual barcodes and whatever information you have for those samples.

Output

  • A bamlist of your samples of interest (for use with angsd)

AND/OR

  • A .clst file you can use for PCA plots (this requires output from pca_calc.sh script as well)

4. SNP Calling (with angsd)

After subsetting to our bamlist to the subpopulation of interest, we can call SNPs using angsd. The binary will need to be installed on your user cluster.

A basic angsd call looks like this:

~/bin/angsd -bam $list -GL 1 -out $output -doMaf 2 -minInd 20 -doMajorMinor 1 -SNP_pval 0.000001 -doGeno 4 -doPost 2 -postCutoff 0.95 -minMaf 0.005

Wrapping this in a shell script is probably the way to go. I have one called ~johnmola/scripts/genoget.sh.

Input

  • a bamlist ($list)
  • an output file name ($output)

Output (can be customized, output below is what the basic angsd call will provide)

  • an arguments file (.arg)
  • minor allele frequency spectrum (.mafs)
  • genotypes (.geno)
Note: They will be .gz files. Just unzip.

Typical angsd commands/calls

angsd calls use a numbering system. For instance, for the -doGeno command, if you want angsd to write the major and minor alleles, you give it the number “1”, if you want posterior probability, you provide “8”. However, if you want both of those values, you provide it 8+1 = “9”

See the angsd wiki for more information. LINK

5. Loci-selection for sibship

In order to run COLONY successfully, we need to select loci that are at least:

  • In HWE
  • In Linkage-equilibrium

Additionally, it’s preferable that we customize the following as well:

  • Select loci from the subset of interest
  • At least 100 loci
  • Less than 25% of individuals are missing a genotype call at that loci
  • Minor allele frequency of 0.1 or higher

In the Geno2Colony.pl file, we will customize the following parameters. Nothing else needs to be edited!! We’ll go through them in detail.

$geno = $ARGV[0];
$list = $ARGV[1];

#SNP Filters
$max_mis = 0.25;
$min_maf = 0.01;
$max_chi = 3.84;

#SNP Drawing
$num = 300;
$dist = 10000;

#Colony
$type = 0;
$adrop = 0.05;
$gerror = 0.05;

Input

$geno is the list of genotypes for all of your individuals. This is the .geno output from angsd.

$list is a header file needed for matching the names to their genotypes. It’s basically just your bamlist, but preferably with human-readable names.

SNP quality filters

$max_mis The maximum number of individuals that can be missing a retained SNP

$min_maf The minimum minor allele frequency for a SNP to be retained

$max_chi Basically assumes HWE if set at 3.84 (don’t change)

SNP drawing

$num Set how many loci you want. The script will randomly pull, but if you set the number really high, it will go until it runs out.

$dist How far apart in base pairs should SNPs be. Typically 10,000. (To ensure linkage equilibrium)

Colony (This input is just needed for COLONY’s header)

$type The type of marker. In this case 0 is the only appropriate answer for SNPs.

$adrop The allele dropout rate. Assuming 0.05 is usually good.

$gerror The genotype error rate. Assuming 0.05 is usually good.

Save the output of this perl script, we’ll feed it to COLONY next. I usually name it with a convention based on the first 5 settings. E.g. The name for the code chunk above would be: projectname_m25_maf01_n300_d10k.DAT meaning the name of the project max missing = 0.25, maf = 0.01, 300 loci, 10000 bp apart.

6. Running COLONY!

Download COLONY here.

We are almost ready to feed our file to COLONY!

We first need to finish the .DAT file that we outputted from the Geno2Colony.pl script. COLONY requires a header and footer of options to be appended onto our offspring genotypes, and then all fed in with one file.

Note: This is different if you use the GUI version of COLONY…but …don’t.

While this step can be automated, I actually prefer some level of interaction here to ensure everything is correct before proceeding. I have saved a typical COLONY header and footer, and then copy and paste them using vim to the beginning and end of the file, respectively. I then ensure the settings are what I want before running COLONY with the folowing command:

srun -t 24:00:00 ./colony2s.ifort.out IFN:./m25_maf10_n500_r1.DAT

Note: For whatever reason, the COLONY software freaks out if you don’t have ./ in the front of your file location.

And…that’s it! You’re done!^^running^^ ^^COLONY^^ ^^at^^ ^^least…^^

X. Quality Checks and Miscellaneous

Y. Glossary of some terms

PLACEHOLDER: I’ve noticed that I use subset, subsample, and subpopulation interchangeably…but to mean different things. Go back through and make clear. Might also be handy to have a glossary to terms that are unique (like bamlist)/weird.

  • bamlist -
  • subpopulation -
  • subsample -
  • subset

Z. Simple version, just scripts and the call

1. Split/align

Split

sbatch -p high -t 24:00:00 ../run_BestRadSplit_PstI.sh SOMM185_R1_ACTTGA.fastq SOMM185_R3_ACTTGA.fastq SOMM185_ACTTGA

Compress

srun -t 24:00:00 gzip *ACTTGA.fastq

Make lists

ls *RA* | sed "s/\.fastq\.gz//g" > listA ls *RB* | sed "s/\.fastq\.gz//g" > listB paste listA listB > list_ACTTGA

Align

sbatch -t 24:00:00 ../run_align.sh list_ACTTGA ../../Mola/B_imp_ref_genome/GCF_000188095.1_BIMP_2.0_genomic.fna

Tidy

mkdir aligned_files filtered_bams slurm_outputs zipped_fastq then mv

Check

ls filtered_bams/ | wc

2. Filter

Remove extension

ls *flt.bam | sed 's/\.bam//g' > sublist

Run sub sample

sbatch -t 24:00:00 sub_sample.sh sublist 100000

Make new bamlist

ls *_100000.bam > _subbamlist

3. Subset

Follow

bamclst_lists.R

4. SNP calling

Edit and run genoget.sh

sbatch -p high --mem 32G -t 72:00:00 ~/scripts/genoget.sh vm_30_bam vm3_30_out

5. Loci selection

Edit and run Geno2Colony.pl

perl ~/scripts/Geno2Colony.pl ../mp_v5_2017_03_16_noclones.geno ../v3_perl_headers > m25_maf01_n75

6. COLONY

Add header and footer

vim

Run COLONY

srun -t 24:00:00 ./colony2s.ifort.out IFN:./m25_maf10_n500_r1.DAT