RESUMEN
Gene regulatory networks control the complex programs that drive development. Deciphering the connections between transcription factors (TFs) and target genes is challenging, in part because TFs bind to thousands of places in the genome but control expression through a subset of these binding events. We hypothesize that we can combine natural variation of expression levels and predictions of TF binding sites to identify TF targets. We gather RNA-seq data from 71 genetically distinct F1 Drosophila melanogaster embryos and calculate the correlations between TF and potential target genes' expression levels, which we call "regulatory strength." To separate direct and indirect TF targets, we hypothesize that direct TF targets will have a preponderance of binding sites in their upstream regions. Using 14 TFs active during embryogenesis, we find that 12 TFs showed a significant correlation between their binding strength and regulatory strength on downstream targets, and 10 TFs showed a significant correlation between the number of binding sites and the regulatory effect on target genes. The general roles, e.g. bicoid's role as an activator, and the particular interactions we observed between our TFs, e.g. twist's role as a repressor of sloppy paired and odd paired, generally coincide with the literature.
RESUMEN
Studies of linkage and linkage mapping have advanced genetic and biological knowledge for over 100 years. In addition to their growing role, today, in mapping phenotypes to genotypes, dense linkage maps can help to validate genome assemblies. Previously, we showed that 40% of scaffolds in the first genome assembly for the Pacific oyster Crassostrea gigas were chimeric, containing single nucleotide polymorphisms (SNPs) mapping to different linkage groups. Here, we merge 14 linkage maps constructed of SNPs generated from genotyping-by-sequencing (GBS) methods with five, previously constructed linkage maps, to create a compendium of nearly 69 thousand SNPs mapped with high confidence. We use this compendium to assess a recently available, chromosome-level assembly of the C. gigas genome, mapping SNPs in 275 of 301 contigs and comparing the ordering of these contigs, by linkage, to their assembly by Hi-C sequencing methods. We find that, while 26% of contigs contain chimeric blocks of SNPs, i.e., adjacent SNPs mapping to different linkage groups than the majority of SNPs in their contig, these apparent misassemblies amount to only 0.08% of the genome sequence. Furthermore, nearly 90% of 275 contigs mapped by linkage and sequencing are assembled identically; inconsistencies between the two assemblies for the remaining 10% of contigs appear to result from insufficient linkage information. Thus, our compilation of linkage maps strongly supports this chromosome-level assembly of the oyster genome. Finally, we use this assembly to estimate, for the first time in a Lophotrochozoan, genome-wide recombination rates and causes of variation in this fundamental process.