Rare Jackpot Individuals Drive Rapid Adaptation in Threespine Stickleback

Résumé fourni par la source

# Notes on scripts 1. We developed a novel approximate likelihood-based method that can be applied to estimate haplotypes of freshwater-adaptive alleles in any Threespine Stickleback genome. This program takes information from multi-SNP loci each with SNPs that have polarized alleles. i.e., for each snp within a mutli-SNP haplotype, we have identified which allele increases in freshwater environment. See the section "Jackpot carriers increased in frequency in Scout Lake" and the section "Genotype calling and validation" in the Methods in the main paper for more details on these loci, and the approximate likelihood method. There are two versions of the program: one takes bams and estimate the genotypes given the read data and the other takes vcf with genotype probabilities (emitted from the imputation program beagle 4.0). The version of the program that takes bams *(fw_caller_V1.py)* takes hours to run, but the version that takes vcf *(fw_caller_V2.py)* is relatively faster. I have provided the two versions in the code folder. These scripts should call the genotypes at each locus as either 0(homozygous marine), 1(heterozygous) and 2( homozygous greshwater). These genotype calls are stored as numpy array and are used for most of the figures in the paper, including: **Figures 1F, 2A, 3, and 5**. In the reproducible run, which should take less than a minute, I feed numpy arrays with the genotype calls for all individuals in all the timepoints which can be found in the folder /data/Genotypes. The numpy arrays were generated with the fw_caller_V1.py. The results from this run will be plots used to generate panel F of Figure 1. The bam files that were used in our study have been uploaded to SRA, PRJNA1231081. I have also included beagle 4.0 imputed vcf for SC2014 as an example to run the fw_caller_V2.py. As we stated in the main paper, both approaches produced similar haplotype genotypes. 2. **Figure 1B,C and D** were plotted with *plot_sexual_mat.py*, which requires the data SC12_13_SL_sex_mat.csv found in the /data folder. 3. We used the functions in the dadi.py module to generate all the site frequency spectra (SFS) and genetic diversity estimates in **Figure 4B and 4C, Figure 7** and the SFS in the supplementary figures. 4. **Figure 5** was generated with *LD_decay.py* 5. **Figure 6** was generated with *relatedness.py* 6. SLiM simulation scripts as *341_independent_loci_simulations.slim* and *341_independent_loci_simulations_remove_jackpots.slim*. The ouput of the simulations were used for plotting **Figure 8**

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.