Search

Courtyard vernacular underground dwellings most commonly called “dig pit houses”, have been for centuries a shelter for the human kind. Despite their noticeable energy saving properties and security enhancing design features, they present several disadvantages related to indoor environment quality such as poor indoor natural daylighting. The dig pit in these typologies is the main design feature that allows natural light into the rooms. In this paper the effect of changing the courtyard’s geometry on the daylighting performance inside the rooms is investigated by the mean of computer simulations using Ecotect-RADIANCE software pack. The findings of this study show that dimensions corresponding to a WI (Well Index) = 0.5 result in better daylighting performance inside the rooms. Moreover, it shows that increasing the number of wall surfaces on the courtyard surface tend to engender a noticeable improvement of the room’s daylighting.

Background Diseases caused by Pseudomonas syringae (Ps) are recognized as the most damaging factors in fruit trees with a significant economic and sanitary impact on crops. Among them, bacterial canker of apricot is exceedingly difficult to control due to a lack of efficient prophylactic measures. Several sources of partial resistance have been identified among genetic resources but the underlying genetic pattern has not been elucidated thus far. In this study, we phenotyped bacterial canker susceptibility in an apricot core-collection of 73 accessions over 4 years by measuring canker and superficial browning lengths issued from artificial inoculations in the orchard. In order to investigate the genetic architecture of partial resistance, we performed a genome-wide association study using best linear unbiased predictors on genetic (G) and genetic x year (G × Y) interaction effects extracted from linear mixed models. Using a set of 63,236 single-nucleotide polymorphism markers genotyped in the germplasm over the whole genome, multi-locus and multi-variate mixed models aimed at mapping the resistance while controlling for relatedness between individuals. Results We detected 11 significant associations over 7 candidate loci linked to disease resistance under the two most severe years. Colocalizations between G and G × Y terms indicated a modulation on allelic effect depending on environmental conditions. Among the candidate loci, two loci on chromosomes 5 and 6 had a high impact on both canker length and superficial browning, explaining 41 and 26% of the total phenotypic variance, respectively. We found unexpected long-range linkage disequilibrium (LD) between these two markers revealing an inter-chromosomal LD block linking the two underlying genes. This result supports the hypothesis of a co-adaptation effect due to selection through population demography. Candidate genes annotations suggest a functional pathway involving abscisic acid, a hormone mainly known for mediating abiotic stress responses but also reported as a potential factor in plant-pathogen interactions. Conclusions Our study contributed to the first detailed characterization of the genetic determinants of partial resistance to bacterial canker in a Rosaceae species. It provided tools for fruit tree breeding by identifying progenitors with favorable haplotypes and by providing major-effect markers for a marker-assisted selection strategy. Electronic supplementary material The online version of this article (10.1186/s12870-019-1631-3) contains supplementary material, which is available to authorized users.

QQ (quantile-quantile) plots of the observed p-values distribution for lgc and bs G and G × Y BLUPs. Non-corrected generalized linear model (glm – green), glm accounting for structure Q (glm.Q – orange), linear mixed model correcting for relatedness K (emmax.K – blue) and both for K and Q (emmax.KQ – brown) output p-values are compared. The 95% confidence interval is indicated in grey. (PDF 387 kb)

Effect of all detected loci using genome-wide multi-locus and multi-variate (mvLMM) mixed models on lgc and bs G and G × Y BLUPs. (A) Manhattan Plots displaying the –log10 (p-values) over physical positions in approximate 100 to 300 kb windows surrounding associations. Significance level determined with eBIC criterion. Detected SNPs were shown with red stars and overlapping candidate genes indicated with blue arrows. As an indication, Bonferroni threshold = −log10(p-value-thr) = 6.10. Pairwise LD heatmaps ( r vs 2 ) $$ {r}_{vs}^2\Big) $$ were drawn within the genomic window around each candidate. (B) Boxplot displaying the allelic effect for the associated marker. For the loci detected with mvLMM on chromosome 7, the allelic effect is represented for both lgc and bs G BLUPs. (PDF 575 kb)

Linkage disequilibrium (LD) decay over physical position in the apricot core-collection. Pairwise r2 values were computed considering markers in a 1 kb-window. Estimated LD decay according to a non-linear regression is represented for each chromosome. Black curves show non-corrected estimates. Curves of corrected r2 estimates with relatedness K (red), structure Q (orange) and both K and Q (blue) are displayed. r2 mean and standard deviation values are indicated. r2=0.10 threshold are represented with dashed horizontal lines. (PDF 2325 kb)

Population Structure (Q) of the apricot core-collection based on data of 21,942 independent SNPs. Genotypic data pruned in order to keep markers with pairwise r2 values

Relation between susceptibility phenotypes lgc and bs G BLUPs and ancestral fractions in the apricot core-collection. Ancestral fractions Q1 to Q3 are displayed as supplementary variables (in blue) in the PCA. The contribution of the first component reached 78.46% of the total variability. (PDF 97 kb)

Time series of degree-days over winter period in lâ Amarine. 2013 to 2016 annual data from November 15th to March 31st. February month is delimited by the red arrows. (PDF 294 kb)