key: cord-0775081-ktdv7sbf authors: Veloso, Artur; Warr, Gregory W.; Browdy, Craig L.; Chapman, Robert W. title: The transcriptomic response to viral infection of two strains of shrimp (Litopenaeus vannamei)() date: 2010-10-28 journal: Dev Comp Immunol DOI: 10.1016/j.dci.2010.10.001 sha: eb0b73f2b5a3a098c7e4107e50b48cb9111705dd doc_id: 775081 cord_uid: ktdv7sbf The extent to which data-intensive studies of the transcriptome can provide insight into biological responses is not well defined, especially in the case of species (such as shrimp) where much physiological and biochemical knowledge is missing. In this study we took a transcriptomic approach to gain insight into the response to viral infection of two strains of the Pacific whiteleg shrimp (Litopenaeus vannamei) that differ in their resistance to Taura Syndrome Virus (TSV). Changes in gene expression in the hepatopancreas following infection with TSV and Yellow Head Virus (YHV) were assessed using a cDNA microarray containing 2469 putative unigenes. The null hypothesis tested was that significant differences between the transcriptomic responses to viral infection of resistant and sensitive strains would not be detected. This hypothesis was broadly rejected, with the most surprising observation being that the baseline (control, unchallenged) sensitive and resistant strains expressed distinguishable transcriptomic signatures. The resistant line was pre-disposed to lower expression of genes encoding viral (and host) proteins. Many of the genes differentiating resistant and sensitive lines are involved in protein metabolism, cellular trafficking, immune defense and stress response, although it was not possible to clearly identify candidate genes responsible for TSV resistance. In contrast to TSV challenge, YSV either failed to perturb the host transcriptome or created a “confused” response that was difficult to interpret. It is axiomatic that the interactions between pathogens and hosts determine the course of an infection, defining survival or death, sickness or health (Kato-Maeda et al., 2001) . In the case of a viral infection this interaction is complex and intriguing because the pathogen will subvert the host's cellular machinery and bring about changes in the cell's function at all levels (Novoa et al., 2005) . Despite this, there are advantages to studying the effects of viral infection on the transcriptome. Using high throughput methods (such as microarrays) it is possible to measure the expression level of thousands of transcripts in organisms fighting a disease condition, thus potentially identifying genes important to the immune response of the organism as well as genes activated or suppressed by the pathogen (Kato-Maeda et al., 2001) . This experimental approach has been undertaken successfully in several organisms, such as human cell lines (Nam et al., 2003) , Atlantic salmon (Jorgensen et al., 2008) , Drosophila melanogaster (Dostert et al., 2005) and Arabidopsis thaliana (Marathe et al., 2004) . The approach can be facilitated by comparison of lineages that show different susceptibility levels to a certain disease, which could point more directly to genes responsible for the resistance to such disease (Ruby et al., 2006; Sarson et al., 2008) . Interest in the immune response of shrimp has been growing due to their worldwide importance in aquaculture and the growing number of emerging pathogens found to affect this group of organisms (Bachere, 2000) . In order to fight such pathogens, the shrimp use a wide range of mechanisms, including hemocytes (Johansson et al., 2000) ; clotting and prophenoloxidase cascade ; anti-microbial peptides (AMP) (Bachere et al., 2004) ; and dsRNA (Robalino et al., 2007a,b) . Some of the most virulent diseases to affect penaeid shrimp are 0145-305X/$ -see front matter © 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.dci.2010.10.001 caused by viral pathogens, including the White Spot Syndrome Virus (WSSV), TSV and YHV (Lightner, 2005) . The Pacific whiteleg shrimp (Litopenaeus vannamei Boone, 1931) is the most cultured shrimp in the world reaching a revenue close to US$7.5 billion (FAO, 2008) . A major reason for the whiteleg shrimp to have become the most cultured shrimp in the world is the existence of both specific pathogen free (SPF) and specific pathogen resistant (SPR) lineages (Wyban et al., 1992; Argue et al., 2002) . Taura Syndrome was first observed in Ecuador and the causative agent (TSV) is a small icosahedral non-enveloped virus with a 10 kb long genome composed of a positive sense single stranded RNA (ssRNA) that belongs to the Dicistroviridae family (Hasson et al., 1995; Bonami et al., 1997; Mayo, 2005) . Despite being first identified in 1991 in southeast Asia infecting the black tiger prawn (Panaeus monodon), Yellow Head Syndrome has the potential to infect L. vannamei and other penaeid shrimp (Limsuwan, 1991; Lu et al., 1994; Lightner, 2005) . The etiological agent is YHV, a rodshaped enveloped virus with a helical nucleocapsid containing a positive sense 22 kb ssRNA genome, belonging to the Roniviridae family (Wongteerasupaya et al., 1995; Nadala et al., 1997; Tang and Lightner, 1999; Mayo, 2002) . Currently there are L. vannamei Taura Syndrome resistant lineages (Argue et al., 2002) , but despite intense work and some evidence for broad based tolerance (Flegel, 2007) there are no SPR lineages for Yellow Head Syndrome. The mechanism of resistance to TSV in L. vannamei is unknown, but the SPR lineage is more susceptible to certain TSV isolates, even though the genetic distance between the isolates is considerably small (deduced amino acid distance from 0 to 7%) (Tang and Lightner, 2005; Srisuvan et al., 2006) . In order to gain knowledge on the resistance to TSV and on the effects of TSV and YHV in shrimp we investigated the transcriptome of TSV resistant and susceptible shrimp when infected with TSV and YHV using a microarray with 2469 L. vannamei putative genes emphasizing immune-related genes (Robalino et al., 2007a,b) . The results of this study were used (1) to test whether significant changes in the transcriptomes of resistant and control shrimp were induced by viral challenge and (2) if the specific genes regulated by viral infection could provide clues to the mechanisms of resistance and susceptibility to viral infections in shrimp. The experiment was conducted using a SPF TSV resistant shrimp strain (R) (Shrimp Improvement Systems Inc., Islamorada, FL) and a SPF TSV susceptible shrimp strain (S) (Kona, Oceanic Institute, HI). The average weight ± standard deviation of the resistant and susceptible shrimp was 2.4 ± 0.84 g and 3.1 ± 0.86 g respectively. The shrimp were kept in an indoor, semi-closed recirculating system composed of 25 transparent 10 L tanks with particulate, biological and UV light filtration, and were allowed to acclimatize for one week prior to the experiment. The shrimp were injected with 60 L of one of three shrimp extracts prepared as described by Prior et al. (2003) : 1. SPF shrimp (diluted 1:10 −4 , w/v); 2 moribund shrimp injected with TSV (diluted 1:10 −5 , w/v); 3 moribund shrimp injected with YHV (diluted 1:10 −4 , w/v). The hepatopancreas of 5 shrimp were sampled from days 0 to 2 and kept in RNAlater ® (Ambion, Austin, TX) and the remaining shrimp were kept in the tanks and mortality counts were taken for 14 days. The mortality curve of the different treatments was statistically compared using the log-rank test (Harrington and Fleming, 1982) as implemented on the R package survival (Therneau and Lumley, 2007; R Development Core Team, 2008) . In order to facilitate the flow of this paper, the shrimp treatment groups will be referred to by acronyms that will be composed of the shrimp strain (R or S) and the pathogen injected TSV, YHV or C for controls. Total RNA was extracted from the hepatopancreas using the RNeasy kit (Qiagen, Valencia, CA) with the on-column DNase step (RNase-Free DNase, Qiagen) following the manufacturer's instructions. Next, 1 g of the extracted RNA was reverse transcribed, in vitro transcribed into antisense amino allyl RNA and labeled with the Cy3 dye using the Amino Allyl MessageAmp II aRNA kit (Ambion). Finally, 10 g of labeled aRNA was hybridized to the microarrays and they were washed and scanned at 70 PMT in a Sca-nArray Express (Perkin Elmer, Boston, MA) as described by Robalino et al. (2007a,b) (Supplementary Material). The microarray slides used in this experiment were composed of two arrays occupying the top and bottom of the slide and with spots printed in the same order. Each array contained duplicates of each DNA spot printed side-by-side (Robalino et al., 2007a,b) . The foreground and background values were calculated using the QuantArray software (Perkin Elmer) and the output files will be uploaded to the NCBI GEO website. The remaining data analyses were carried out using a set of packages from the open source software R and BioConductor (Gentleman et al., 2004; R Development Core Team, 2008) , with an emphasis on functions from the package limma (Smyth, 2005) . The data were initially background corrected using the normexp model which uses maximum likelihood to calculate the parameters to fit a convolution of a normal and an exponential distribution (based on background and foreground respectively) and from that estimate it calculates the expected foreground (Ritchie et al., 2007) . As the microarrays used were in-house printed (Robalino et al., 2007a,b) and batches of amplicons presented differences in purity, a print-order loess normalization was applied (Smyth and Speed, 2003) . Next, the aroma.light package was used to perform a quantile normalization using a smooth spline that transforms the data to fit an empirical distribution calculated based on the average expression between all arrays (Bengtsson et al., 2008) . Genes for which, the 4 replicates did not show intensity higher than three times the average intensity of the non-shrimp controls (16 Tursiops truncatus genes printed in duplicate in both arrays present in each slide) in at least 10% of the slides in the experiment were disregarded. Lastly, the data were variance stabilization normalized using the vsn package (Huber et al., 2002) . To examine the capacity of our microarrays to discriminate the various treatments we employed a variation of the Artificial Neural Network (ANN) strategy we have described in detail elsewhere (Chapman et al., 2009) . Briefly this approach makes pairwise comparisons of treatment groups (regardless of day of sampling) using all the genes passing the filtering process described in Section 2.3.1 with 5 rounds of ANN training and random sampling of the records (individuals) in each group for each training session. The samples not used in the training session were used as a cross validation set. Following the initial training sessions the ANN models with the highest R 2 were then used to compute the sensitivities of the binary classifications to changes in gene expression. The 250 genes with the highest sensitivities for each comparison were then used in a second round of ANN training accomplished as before but this time with 20 rather than 5 training sessions. To evaluate the power of this reduced set of genes to discriminate treatment groups, Table 1 Area under the ROC curves for comparisons of shrimp treatments using the 250 genes that were most important (of the 2444 total) to the particular comparison. S = sensitive to TSV, R = resistant to TSV, C = control, TSV = challenged with TSV, YHV = challenged with YHV, NA = not applicable. Values in bold are NOT statistically significant at p < 0.01. we employed receiver operating characteristic curves (ROC) as described in Chapman et al. (2009) . Combining treatment groups regardless of day of collection is less than optimal. However, it was necessary, as our initial analysis (which employed typical linear approaches, e.g. control vs. treatment on day 1), while interesting, did not identify with statistical significance many genes. Further this approach does not really address the most important questions: Whether individual gene expression profiles can collectively discriminate resistant versus sensitive lineages and whether these lineages respond differently to viral challenges. Combining data regardless of day of sampling is not a desirable approach, but it has been our experience that ANNs do not perform well with sample sizes less that 10. We did explore the alternative (comparing each treatment day) and found the results less than compelling. In other words, combining the data regardless of day introduces noise due to time delays in the genomic response, but boosts the signals due to larger sample size. At the end of the 14 days of observation the different group mortalities were very similar to the expected. The RTSV groups showed different mortalities to the STSV group (p-value = 1.53 × 10 −6 ) and to the SC and RC groups (p-value = 4.89 × 10 −2 and pvalue = 1.74 × 10 −3 respectively). The mortality of all treatment groups was extremely similar at day 1 (∼7%). At day 2, however, the mortality in RYHV and SYHV (69.6% and 90% respectively) was much greater than the one seen in RTSV and STSV (7.4% and 34.6% respectively) (Fig. 1) . The differences in mortality at day 2 need to be taken in consideration when trying to extract biological information from the microarray experiment. While the YHV infected animals at day 2 were in a final stage of the disease, the TSV infected animals were less sick, such that most of the mortality happened on the following days. It is possible that the differently expressed genes in the YHV treatments at day 2 are more related to the death process, while in the TSV treatments there might be more genes related to immune defense. In Table 1 we present the results of the ROC curves that report the discrimination between the transcriptomic signatures of the control and experimental shrimp groups. These values can be considered as the probability of correctly assigning individuals to the appropriate class. The first comparison was between the control (C, unchallenged) transcriptomes of the shrimp lines resistant (R) and sensitive (S) to TSV infection. The RC versus SC comparison in Table 1 shows that the probability of correctly assigning a resistant or susceptible individual to the correct group is better than 98%; this value differs from random allocations (50:50) by a highly significant (p < 0.001) margin. We interpret this result to suggest that resistance to infection by TSV may be associated not only with the induction of genes by infection, but by a pre-existing pattern of gene expression reflecting a relatively resistant constitutive state. Supplementary Table S1 presents lists of genes that lead to this distinction between the transcriptomic signatures of the RC and SC groups, as well as the fold differences in their mean expression. The numbers in the column should be considered baseline differences between the strains. In this gene list more than half of the genes are unknowns, reflecting our ignorance of the functions of most genes in shrimp. However, of the annotated genes nearly all are involved in cellular trafficking and motility, protein synthesis, energy production or control of energy production, growth, and lipid metabolism/transport. It is significant that the bulk of the known genes, are involved in protein synthesis and cellular trafficking and are down regulated in the TSV resistant strain. The down regulation of energy and growth related genes in the resistant line is consistent with the results of Argue et al. (2002) who reported that the TSV resistant line grew more slowly than control lines and that there was a negative correlation between growth and TSV resistance. The results present here provide a molecular underpinning for this observation. As the protein synthesis and cellular trafficking machinery are precisely the components of cellular activity that are essential for viral replication, it is logical that down-regulation of these systems would inhibit viral induced mortality as was evident in Fig. 1 . The translationally controlled tumor protein (TCTP), also called fortilin (Tongsnunt et al., 2008) is slightly upregulated in resistant lines, This protein has been shown to play a role in protecting shrimp from white spot syndrome virus (WSSV) (Tongsnunt et al., 2008) . The inhibition of the transactivation response (TAR) RNA binding protein (TRBP) has been shown to inhibit replication of HIV in human cells, presumably through the action of TRBP on protein kinase R activation which is essential for HIV replication (Christensen et al., 2007) . It would, however, be premature to suggest that the mechanism of TSV resistance in shrimp is analogous to HIV resistance in human cells, although genes that play a major role in trafficking mRNA to the rough endoplasmic reticulum for translation, are down regulated in the resistant line. Three transcription initiation factors and one elongation factor are also down regulated in the resistant line. This is significant in that TSV is a ss(+) RNA virus whose genome effectively behaves as mRNA. This suggests that the resistant line has a reduced capacity for translating the TSV messages (as well as it own messages) compared to the sensitive line. Overall, comparisons of transcriptomes from unchallenged resistant and sensitive lines suggest that the resistance to TSV is at least in part due to a predisposition of the resistant lineage to slow viral replication. The transcriptomic changes resulting from TSV infection in the resistant line provided some discrimination of the treatment groups (AUC = 0.73, RC vs. RTSV, Table 1 ) and while this classification precision was significantly greater than random expectations, it was substantially less (p < 0.001, one tailed z-test of the AUC's) than the correct classification level for the resistant versus sensitive control comparisons. The implication is that resistance to TSV may involve some transcriptomic adjustments, but it reinforces the notion (discussed above) that preprogramming is a major feature of resistance to TSV infection. The genes discriminating the transciptomic responses of the resistant control shrimp versus resistant shrimp challenged with TSV include a long list of energy metabolism related proteins, some secretory systems (sec61, sec24, and other translocators), initiation factors, protein synthesis, and cytokinesis genes (actin and related factors including spaghetti squash and spindle pole body), which are slightly more represented in the up regulation than down regulation classes. While it is tempting to infer some role of these genes in viral responses, it has been our experience and that of others (Chapman et al., 2009; Gracey, 2007; Gracey et al., 2008; Hofmann and Place, 2007; Place et al., 2008; O'Donnell et al., 2009; Evans and Somero, 2008) that energy metabolism, cytokinesis and protein synthesis genes are general indicators of organismal stress (including chemical, pH, temperature, land use and disease) in a variety organisms including oysters, mussels, shrimp, fish and marine mammals. Thus the bulk of the known genes that discriminate viral infection from the control probably reflect general stress responses rather than a specific response to viral challenge. Genes down regulated by TSV infection were dominated by unknowns, particularly those with the higher fold changes in expression. Absent from this gene list is the TRBP gene mentioned above, which indicates that this gene was not a major contributor to discrimination of the impact of TSV on the resistant line. The discrimination of the transcriptomic response of the resistant line to YHV infection (AUC = 0.86, Table 1 ) was significantly better (p < 0.001) than the discrimination (from the unchallenged controls) of the transcriptomic response of the TSV resistant shrimp to TSV infection. This implies that YHV elicits a more pronounced change (or at least one more easily identified) in the transcriptome than the changes induced by TSV in the resistant line. This is in keeping with what one might expect from the relative mortalities shown in Fig. 1 . In comparing the YHV challenged individuals with the controls it was evident that fortilin gene expression was up regulated while the fortilin binding protein gene was strongly down regulated Tongsnunt et al. (2008) reported a strong positive correlation of these genes in the response of Peneaus monodon to WSSV infection. The data presented here suggest the action of these genes in response to general viral infection may be more complicated, especially as, interestingly, the fortilin protein was down regulated by TSV infection. Previous studies have implicated components of the RNAi pathway, especially the Argonaut and Sid families, in viral responses of shrimp (Robalino et al., 2007a,b; Labreuche et al., 2010) but these genes were not printed on the microarray used in these studies. Unlike the resistant lineage, the TSV sensitive line differed dramatically in the transcriptomic response to TSV and YHV. The response to TSV could not be distinguished from random effects, as the area under the curve of 0.59 was not statistically significant (Table 1 ). This suggests that TSV infection in the sensitive line does not initiate large changes in the transcriptome and that the virus manages to evade discovery by the antiviral defense mechanisms. In contrast the transcriptomic response to YHV was highly discriminated (area under the curve 0.91, p < 0.001), indicating that YHV initiates a substantial modification of the transcriptome. YHV infection influences the expression of a variety of ribosomal, translocator, elongation and initiation factors as well and energy metabolism genes. However, there does not appear to be a consistent pattern of up or down regulation as genes in all of the classes are found in both up and down states relative to the control. It is as if YHV is not escaping notice, but creating confusion in the host as to what the appropriate response should be. One gene of interest is endoribonuclease U (endo U), which is widely distributed in metazoans and involved in intron excision (Renzi et al., 2006) . A similar protein has been discovered in the Nodovirales order and is essential for viral translation and replication of the coronavirus which causes severe acute respiratory syndrome (SARS) in humans (Ivanov et al., 2004) . As YHV is in the related Roniviridae family, it is possible that the signal on the microarrays is confounded by a combination of endogenous shrimp response and viral genes. We doubt this is the case, as 18 kb of the YHV genome has been sequenced (Senapin et al., 2010) and no match to endo U is apparent. Overall, the sensitive line does not provide a great deal of information regarding the mechanisms of shrimp response to viral infection other than the tantalizing notions that TSV manages to escape detection (an ideal ability for any virus, Hay and Kannourakis, 2002 ) and YHV appears to create confusion which may be next best choice. In this report we have examined the transcriptomic response of TSV sensitive and resistant lines to two viral pathogens, concluding that TSV resistance is to some degree preprogrammed in the resting state of the resistant shrimp. However, there is likely also to be some genetically determined response that also contributes to resistance. In the resistant line we see some indication of genes that may be involved in resistance to TSV including genes with known capacity to limit apoptosis and viral amplification. However, we should note that inhibiting apoptosis is the Sword of Damocles. In the early stages of infection apoptosis favors the host by destroying the cellular machinery necessary for viral replication. Once the infective virions have been produced, apoptosis favors the virus by encapsulation of the virions in host vesicles, which facilitates transmission to other cells (Hay and Kannourakis, 2002) . In other words, timing, while it may not be everything, is important. As significant TSV resistance can be generated in as little as one generation, it is likely that a small number of genes are responsible for the resistance. It is therefore unlikely that the large number of genes discussed here are directly responsible for resistance, but rather some master regulator of their expression amenable to selection is at the heart of the response. Further, we can, on the basis of this study, dismiss a large number of known genes as likely candidates underpinning resistance, as they are known to be involved in general stress response in shrimp and other organisms. We should also dismiss at this point the fortilin, fortilin binding protein, TRBP and others. Our reasoning is that (1) these are effectors and not gene regulators, and (2) there are too many of them to respond to selection in a single generation unless they are tightly linked. We view it as more likely that resistance to TSV is conferred by some combination of genes of unknown function and/or genes not represented on this array. Selective breeding of Pacific white shrimp (Litopenaeus vannamei) for growth and resistance to Taura Syndrome Virus Shrimp immunity and disease control Insights into the anti-microbial defense of marine invertebrates: the penaeid shrimps and the oyster Crassostrea gigas Estimation and assessment of raw copy numbers at the single locus level Taura syndrome of marine penaeid shrimp: characterization of the viral agent A transcriptomic analysis of land use impacts on the oyster Crassostrea virginica Small interfering RNAs against TAR RNA binding protein TRBP, a dicer cofactor, inhibit human immunodeficiency virus Type 1 long terminal repeat expression and viral productions The Jak-STAT signaling pathway is required but not sufficient for the antiviral response of drosophila A microarray-based transcriptomic time-course of hyper-and hypoosmotic signaling events in the euryhaline fish Gillichthys mirabilis: osmosenors to effectors Global Aquaculture Production (online query Update on viral accommodation, a model for host-viral interaction in shrimp and other arthropods Bioconductor: open software development for computational biology and bioinformatics Rhythms of gene expression in a fluctuating intertidal environment Interpreting physiological responses to environmental change through gene expression profiling A class of rank test procedures for censored survival data Taura syndrome in Penaeus-Vannamei -demonstration of a viral etiology A time to kill: viral manipulation of the cell death program Genomics-enabled research in marine ecology: challenges risks & pay-offs. Invited review Variance stabilization applied to microarray data calibration and to the quantification of differential expression Major genetic marker of nidoviruses encodes a replicative endoribonuclease Crustacean haemocytes and haematopoiesis Gene expression analyses in Atlantic salmon challenged with infectious salmon anemia virus reveal differences between individuals with early, intermediate and late mortality Microarray analysis of pathogens and their interaction with hosts Non-specific activation of antiviral immunity and induction of RNA interference may engage the same pathway in the Pacific white leg shrimp Litopenaeus vannamei Biosecurity in shrimp farming: pathogen exclusion through use of SPF stock and routine surveillance Histology and ultrastructure reveal a new granulosis-like virus in penaeus-monodon affected by yellow-head disease Infection of the Yellow Head Baculo-Like Virus (Ybv) in 2 Species of Penaeid Shrimp Penaeus-Stylirostris (Stimpson) and Penaeus-Vannamei (Boone) Study of Arabidopsis thaliana resistome in response to cucumber mosaic virus infection using whole genome microarray A summary of taxonomic changes recently approved by ICTV Changes to virus taxonomy Detection of yellowhead virus and Chinese baculovirus in penaeid shrimp by the Western blot technique Expression of interferon inducible genes following Hantaan virus infection as a mechanism of resistance in A549 cells Virus factories: associations of cell organelles for viral replication and morphogenesis Predicted impact of ocean acidification on a marine invertebrate: elevated CO2 alters response to thermal stress in sea urchin larvae Gene expression in the intertidal mussel Mytilus californianus: physiological response to environmental factors on a biogeographic scale Controlled bioassay systems for determination of lethal infective doses of tissue homogenates containing Taura syndrome or white spot syndrome virus R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing The structure of the endoribonuclease XendoI: from small nucleolar RNA processing tro severe acute respiratory syndrome coronavirus replication A comparison of background correction methods for two-colour microarrays Insights into the immune transcriptome of the shrimp Litopenaeus vannamei: tissue-specific expression profiles and transcriptomic responses to immune challenge Double-stranded RNA and antiviral immunity in marine shrimp: inducible host mechanisms and evidence for the evolution of viral counter-responses Transcriptional profiling reveals a possible role for the timing of the inflammatory response in determining susceptibility to a viral infection Transcriptional analysis of host responses to Marek's disease virus infection in genetically resistant and susceptible chickens Impact of yellow head virus outbreaks in the whiteleg shrimp Limma: linear models for microarray data Normalization of cDNA microarray data Comparison of four Taura syndrome virus (TSV) isolates in oral challenge studies with Litopenaeus vannamei unselected or selected for resistance to TSV The proPO and clotting system in crustaceans Phylogenetic analysis of Taura syndrome virus isolates collected between 1993 and 2004 and virulence comparison between two isolates representing different genetic variants A yellow head virus gene probe: nucleotide sequence and application for in situ hybridization Survival: Survival Analysis, Including Penalised Likelihood The role of Pm-fortilin in protecting shrimp from white spot syndrome virus (WSSV) infection Yellow-Head Virus of Penaeus-Monodon Is an Rna Virus Development and commercial performance of high health shrimp using specific pathogen free (SPF) broodstock Penaues vannamei This research was supported by the USDA CSREES U.S. Marine Shrimp Farming Program and by a Marine Genomics fellowship support from the College of Charleston to A.V. and was conducted at the Hollings Marine Laboratory, Charleston, SC. This paper has a contribution number 676 for MRD SCDNR and 54 for Marine Biomedicine and Environmental Sciences Center. Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.dci.2010.10.001.