Via a genome scan analysis, this study, based on the daughter design described by Soller and Genizi , has identified five QTL influencing parasite resistance traits on four sheep autosomes. Considering that two independent traits were analysed (according to a principal component analysis performed with the SAS® package ; results not shown), the numbers of tests in our experiment expected by chance alone to be significant at the 5% genome-wise and chromosome-wise level are 0.13 and 2.6, respectively. We identified one and four significant associations in our across-family analysis for these respective significance levels, providing evidence in favour of genuine segregating QTL for parasite resistance traits in the studied population of Churra sheep.
By adapting the method proposed by Weller et al.  to our experimental conditions (e.g., the number of ewes and families analysed, marker density and marker informativeness), we estimated that the power of this experiment to detect a QTL with two alleles that occur with equal frequency and influence a trait with a heritability of 0.20 varied between 16% (0.3 phenotypic SD units) and 42% (0.5 phenotypic SD units) according to the magnitude of the allelic substitution effect that we considered. This estimation was performed assuming a type I error rate of 0.05 and 10% recombination between a marker and the QTL. Hence, we should take into account the fact that the low number of animals analysed in the regression analysis had an important negative influence on the statistical power of the experiment, and that a substantial proportion of genuine segregating QTL, especially those with small effects, may not have been identified by the across-family regression analysis performed. Therefore, we suggest that some of the other regions that were identified at a lower significance level in the across-family analysis might represent genuine QTL segregating in individual families. Some of these weak associations, e.g., QTL identified at the 10% chromosome-wise significance level for Peps on chromosomes 1, 2 and 24, IgA on chromosomes 9 and 13, and LFEC
1 on chromosome 26 (data not shown), might be confirmed if additional animals were to be included in the analyses.
The lack of coincidence among the QTL identified for the different traits analysed here supports our previously mentioned hypothesis that the traits studied may represent different aspects of the host-parasite interaction during infection. It is possible that the QTL detected for IgA and Peps could be related to the early response to incoming larvae (i.e., hypersensitivity reactions), whereas the QTL for faecal egg counts may be associated with the ability to avoid the development of adult parasites. This agrees with the observations reported by Davies et al. , who did not find any coincident QTL between parasitic traits and IgA activity. The lack of coincidence between the QTL influencing LFEC
0 and LFEC
1, although intriguing, agrees with certain differences observed regarding the correlations between these two traits and the serum indicator traits . As suggested in that work, this could be related to the limited sample period between the faecal egg counts, which could indicate that LFEC
1 is a better indicator of the initial immune response triggered by larvae at the beginning of infection.
On the other hand, the allelic substitution effects estimated for the QTL reported herein are likely to be overestimated as a result of the low power of the experiment at the sire-marker level. As shown by Lynch and Walsh , the lower the power, the more the effects of a detected QTL are overestimated. Hence, the genuine QTL effects are likely to be much smaller. This result would be in accordance with the work of Houle et al. , who suggested that parasite resistance is likely to be controlled by several loci and, therefore, may receive a strong mutation input, which generates genetic variation. This agrees with the complexity of the physiological processes that lead to nematode resistance .
In order to compare our QTL analysis results with chromosomal regions previously identified in sheep in relation to parasite resistance traits, we consulted the Sheep Quantitative Trait Loci (QTL) database  and other reports available in the literature. We found that some previously published QTL are coincident with the results reported herein. It is worth noting, however, that most of the QTL mapping studies targeting parasite resistance traits in sheep have typically used experimentally challenged animals, and that the parasite species considered vary between studies. In addition, most of the previously reported studies consider parasite resistance traits measured in young animals, mainly meat production lambs.
Marshall et al.  recently reported a QTL on chromosome 1 for Haemonchus contortus faecal egg count in 13-month-old Australian sheep. This QTL is close to the marker ADMST4, which maps within the flanking interval of the chromosome 1 QTL reported here for LFEC
1. At the proximal end of the same chromosome, within the marker interval EPCDV010-ILSTS044, Díez-Tascón et al.  reported a within-family QTL for faecal strongyle egg count and an across-family significant QTL for adult T. columbriformis recovered from the gastric contents of outcrossed lambs at slaughter. These significant associations co-localise with the position of the chromosome 1 QTL influencing IgA that was identified in Churra sheep in our analysis.
On chromosome 6, Beh et al.  reported a genome-wise significant QTL for faecal T. columbriformis egg count in lambs after primary challenge. This QTL was confirmed to have a chromosome-wise significance following a secondary challenge and mapped to the interval between markers MCMA22 and MCM214. According to the latest version of the Australian Sheep Linkage Map (v 4.7) , the first of these two markers is 16 cM distal to CSN3 (male map), one of the markers flanking the genome-wise significant QTL identified by our across-family regression analysis.
On chromosome 14, Davies et al.  reported three QTL related to Nematodirus egg count in Scottish blackface lambs that were located in the last third of the chromosome, whereas the QTL for LFEC
0 that we identified mapped to the centromeric end of chromosome 14.
Considering the low resolution of the preliminary genome scans that have been reported thus far regarding QTL position, some of these coincidences might indicate common underlying loci affecting parasite resistance traits. However, this possibility should be confirmed with further studies. Taking into account the high degree of variation between different experiments due to factors such as the type of parasite exposure (natural or artificial challenge), the parasite species, the phenotypic indicators and the breeds of sheep studied, the identification of non-coincident QTL in different experiments may suggest the existence of complex host-parasite relationships that have unique features that depend on the host-parasite combination.
Curiously, our analysis did not find any significant association within two of the regions for which consensus has been found in different studies. These are the regions close to IFNG on chromosome 3 [7, 34] and the histocompatibility complex (MHC) region on chromosome 20 [7, 35, 36]. This discrepancy may be explained by the fact that the studies that found significant associations in these two regions were focused on lambs, whereas our study considered adult ewes. Marshall et al.  reported an important age and/or immune status specificity of the QTL for resistance to Haemonchus contortus that they identified in Australian sheep. This specificity is based on the low overlapping levels observed for the QTL that influenced the faecal egg counts measured in animals 6 and 13 months of age. This kind of age-specific mode of action could apply to most parasite infections, which would provide support at the genetic level for the hypothesis suggested by Stear et al.  that describes the different mechanisms controlling GIN parasite infections in lambs (antibody response) and adult sheep (hypersensitivity reaction). Also, Balic et al.  suggested that the genes that control key mechanisms preventing the establishment of worms in primary infections are different from those involved in subsequent infections. This idea is based on the different pathways that are involved in innate and acquired resistance. However, this hypothesis is challenged by the fact that overall immunity has been successfully achieved through selection for acquired resistance rather than via resistance to primary exposure to worms . All these observations highlight the complexity of parasite resistance and the difficulty of completely understanding the genetic architecture of the physiological mechanisms underlying resistance as well as resilience. As mentioned by Dominik , consistency in protocols, experimental materials and analysis approaches would facilitate the generation of phenotypic information that would help to increase our knowledge on this topic.