Additional file 1 of Using deep learning to identify recent positive selection in malaria parasite sequence data

Autor: Deelder, Wouter, Benavente, Ernest Diez, Phelan, Jody, Manko, Emilia, Campino, Susana, Palla, Luigi, Clark, Taane G.
Rok vydání: 2021
DOI: 10.6084/m9.figshare.14782785.v1
Popis: Additional file 1: Table S1. Simulation parameters for the data generation using SFS_Code software. Table S2. Performance of Convolutional Neural Network (CNN) model structures on simulated datasets. Table S3. Sample origin by geographic location. Table S4. The 1,125 high-quality P. falciparum isolates used in this study. Table S5. The 368 high-quality P. vivax isolates used in the study. Table S6. Plasmodium falciparum loci identified by DeepSweep (DS; with >3 SNPs). Table S7. Plasmodium vivax loci identified by DeepSweep (DS; with >3 SNPs). Table S8. Plasmodium falciparum loci with the most iHS and Rsb hits. Table S9. Plasmodium vivax loci with the most iHS and Rsb hits. Figure S1. The creation of haplo-images. Figure S2. Workflow. Figure S3. Exemplar images of simulated isolates undergoing different types of sweeps or neutral evolution. Figure S4. Model performance on simulated datasets. Figure S5. Distribution of the minor allele frequencies across the SNPs. Figure S6. Model performance for Plasmodium falciparum and P. vivax on training and validation datasets. Figure S7. Relationship between -log10 p-value of Rsb hits and number of DeepSweep hits.
Databáze: OpenAIRE