Skip to main content
NCBI home page
Poultry Science logoLink to Poultry Science
. 2026 Jul 29;105(11):107522. doi: 10.1016/j.psj.2026.107522

The physiological drivers of commercial broiler chicken performance uncovered with transcriptomics analyses and a novel pathway activation method

LF Romero a,⁎, CEC Blue b, DA Fenster b, D Ritter c, L Payling a, K Patel a, RA Dalloul b
PMCID: PMC13487288  PMID: 42580270

Abstract

Factors contributing to sub-optimal animal performance are often subtle and difficult to identify. This proof-of-concept study generated transcriptomics profiles from chickens’ jejunum and liver tissues to understand differences among broiler flocks with varying final growth performance. Two pathway analysis methods were used and compared, including an enrichment method (Gene Set Enrichment Analysis, GSEA) and a topology-based, Quantitative Pathway Activation method (QPA, Biofractal, Portugal). Straight run Ross-708 broilers were selected from four different farms within a commercial broiler integration that differed in historic FCR at 28 days of age. Farms were retrospectively assigned to a performance category (High Performing (HP) n = 2 or Low Performing (LP) n = 2) according to performance at the end of the cycle (63 d). Final FCR were 1.901 and 1.914 for the HP farms and 1.923 and 1.951 for the LP farms. Final BW were 4,246 and 4,309 g/bird for the HP farms and 4,096 and 4,188 g/bird for the LP farms. Birds from HP farms had lower crypt depth (270 µm vs 381 µm) and greater villi length to crypt depth ratio (3.95 vs 2.69) than LP farms (P < 0.05). The GSEA and QPA analyses showed that LP birds had activation of immune function and inhibition of nutrient metabolism in the jejunum and liver compared to birds from HP farms. QPA results suggested a relatively lower nutritional state in the LP birds compared to HP birds including inhibited energy, lipid, and amino acid metabolism in the liver, and inhibited digestion and absorption processes in the jejunum. Both the GSEA and QPA data suggested that LP birds were exposed to cellular stress. Whilst GSEA recorded this stress as pathways related to heat stress, the QPA method detected hypoxia and oxidative stress as sources of cellular stress that were directly associated with reduced final BW in LP birds. In conclusion, transcriptomics data from commercial broiler farms demonstrated potential to provide in-depth insights into the health and nutrition of broiler chickens. Quantitative pathway activation methods and improved functional annotation of biochemical pathways demonstrated high potential for field applications such as detection of limitations for performance and evaluating the effectiveness of nutritional interventions.

Keywords: Transcriptomics, Broiler, Performance, Histology

Implications

Transcriptomics data analysis from commercial broiler farms provided insights that helped to explain the deviation in production performance in a commercial setting. The methods of pathway analysis were important, with the Quantitative Pathway Activation method showing more specific insights that were easier to interpret from an animal nutrition and health perspective compared to the commonly used Gene Set Enrichment Analysis method of pathway enrichment. This information could be used to optimize nutritional strategy including diet formulation and feed additive inclusion. There is potential for field application of these methods to enable precision nutrition and the optimization of commercial broiler performance.

Alt-text: Unlabelled box dummy alt text

Introduction

There is a substantial gap between the genetic potential of chickens for productivity and their performance in commercial conditions. Sub-clinical challenges limit the productivity of broiler chicken flocks, yet their subtle physiological changes are often difficult to detect (Stanley et al., 2014). Moreover, the intricate interplay between nutrition and health demands a comprehensive understanding of the underlying physiological mechanisms governing broiler health and performance (Leeson and Summers, 2005).

Modern omics technologies present a promising option for enhanced flock performance through precision nutrition and health management. Among these, gene expression analysis, particularly mRNA sequencing (RNA-seq), is popular for its ability to explore the molecular mechanisms of the host in a relatively unbiased manner (Huang et al., 2009). However, challenges such as data complexity, sample variability, and computational demands underscore the need for robust analytical frameworks to extract meaningful biological insights (Conesa et al., 2016).

Among the commonly employed methods of gene expression analysis, differential gene expression analysis and pathway enrichment, notably Gene Set Enrichment Analysis (GSEA), have emerged as common tools in the RNA-seq repertoire (Reimand et al., 2019; Subramanian et al., 2005). These methods identify genes and pathways that are up- or down-regulated in a group of interest compared to a reference population. However, the pathway catalogues on which pathway enrichment methods rely are incomplete, generic across species and tissues, and difficult to interpret for applied scientists because they are annotated primarily for molecular biology scientists and basic science research.

This paper presents a proof of concept for the use of gene expression analyses as tools for understanding commercial broiler flock performance and informing health and nutrition strategies. Two pathway analysis methods are evaluated, including the commonly used GSEA method (Subramanian et al., 2005), as well as the Quantitative Pathway Activation method (QPA, Biofractal Portugal) complemented by a phenotypic association method (Biofractal, Portugal), which apply custom functional pathway catalogues, optimized pathway activation algorithms, and machine learning to relate pathways to animal phenotypes, with the view that these methods may improve the insight generation and application of gene expression data in animal sciences.

Materials and methods

Experimental design

Four flocks of straight run Ross-708 broilers of the same age (28 ± 1 d) were selected from four farms in the same complex of a commercial broiler integration in the USA, with a wide variation of recent historic FCR. At d 28, 16 healthy birds (8 males and 8 females), were randomly selected from each farm for sampling, proportionally representing each house in each farm. Four visibly healthy chickens (two males and two females) of average BW were selected, taken randomly using a zig-zag walk in each of four houses per farm. At the end of the broiler production cycle at 63 d for all farms, adjusted FCR was calculated (Table 1), and farms were categorized a-posteriori into HP (farms 2 and 3) or LP farms (farms 1 and 4). All flocks were processed at the same age (63 d) and differed only in final live weight (4096–4309 g). We standardised the FCR of each flock to that of the lightest flock (Farm 4, 4096 g) to allow comparison at a common weight. The correction factor of 0.01 FCR units per 45 g live weight was taken from the slope of the cumulative FCR-versus-body-weight relationship at the upper end of the Ross 708 growth curve in the breed performance objectives (Aviagen, 2019), where FCR rises by approximately 0.019 units per 85 g of gain (≈0.01 per 45 g) between days 54 and 56. Each flock's FCR was then adjusted as: adjusted FCR = measured FCR − [(flock BW − reference BW)/45 × 0.01.

Table 1.

Final performance of four commercial broiler flocks.

Metric Farm 1 Farm 2 Farm 3 Farm 4
Housed birds (#) 113,100 81,700 154,400 78,000
Slaughter age (d) 63 63 63 63
Mortality at 63d (%) 6.54 4.32 4.6 5.38
Final average BW (g/bird) 4,187 4,246 4,309 4,096
Final FCR (g/g) 1.953 1.901 1.914 1.923
Adjusted FCR1 (g/g) 1.933 1.868 1.867 1.923
Performance Category Low Performing High Performing High Performing Low Performing
Open in a new tab
1

FCR was adjusted to the weight of Farm 4 based on the changes of FCR at different BWs for that age (0.01 for 45 g) in the breed’s Performance Guidelines (Aviagen, 2019).

The effect of performance category on histology variables and mRNA expression analyses was assessed using individual birds as experimental units. The experiment was performed in accordance with relevant guidelines and regulations and was approved by a company animal care and use committee.

Commercial flocks

Flocks were on a no-antibiotics ever (NAE) program and came from the same hatchery plant, where they received a coccidiosis vaccine after hatch. All birds were inovo vaccinated with HVT+IBD vector and SB1 vaccines and all flocks were subject to the same vaccination program. Flocks were fed the same all-vegetable diets. All flocks were housed within two days of each other in June 2021 and all were later sampled on the same day. The farms had initial flock sizes between 78,000 and 154,400 birds (Table 1). At slaughter (63 d), final BW, FCR, and mortality were calculated for each farm. FCR was BW adjusted based on farm 4 (0.001 for 45 g). Final performance results supporting the performance category allocation are presented in Table 1.

Sampling

When birds were 27 or 29 d of age (Table 2), a total of 16 healthy, randomly selected birds per farm (8 males and 8 females) were weighed and euthanized by cervical dislocation. Sampling was performed proportionally from all houses within a farm and in different locations within each house, covering the entire house. Within 1 min of euthanasia, samples of 0.5 by 0.5 cm from the bottom of the right lobe of the liver and a cross section from the middle loop of the jejunum were collected in pre-labelled cryo-vials containing RNA later for stabilisation of the tissue RNA in preparation for mRNA sequencing and kept at refrigeration temperature prior to arrival at the laboratory, and storage at −20°C. An additional cross section sample of 0.5 cm of jejunum, 1.5 cm anterior to Meckel’s diverticulum was collected in pre-labelled cryo-vials with NOTOXhisto (Scientific Device Laboratory, Des Plaines, IL, USA) for histological analyses. Lastly, necropsy was performed by a single qualified veterinarian, coccidiosis lesions were visually scored (1-4) and evidence of necrotic enteritis was recorded.

Table 2.

Summary of sampling information by farm, live BW and average daily gain of birds sampled.
Probability2
Performance Category1 High Performing Low Performing High Performing Low Performing Farm Performance Category
Farm Farm 2 Farm 3 Farm 1 Farm 4
Sampling age (d) 27 29 27 29 28 28 — —
Flock mortality at sampling age (%) 2.33 2.49 2.84 1.95 2.43 2.48 — —
Birds sampled (#) 16 16 16 16 32 32 — —
Males/Females (#) 8/8 8/8 8/8 8/8 16/16 16/16 — —
Live BW at sampling (g/bird) 1,179a,b 1,307a 1,114b 1,239a,b 1,248a 1,188b 0.001 0.1
ADG from hatch to sampling (g/bird) 44.7 45.6 41.5 43.2 45.2a 42.4b 0.1 0.026
Open in a new tab
a,b

Means with different superscripts differed at P < 0.05. Individual birds were considered experimental units.

1

Performance Categories were established a-posteriori, based on adjusted FCR at 63 d of age.

2

Statistical differences between Farms and Performance Categories were independently evaluated.

Laboratory analyses

Histology slides were examined using blind labels, without prior knowledge of the assigned performance category. Representative intact villi and crypts were measured (Fig. 1) as a small cluster using a KR406 10 × 10 grid reticule on a Leica DM3000 microscope, calibrated to micrometres with AmScope AM400 calibration slide. RNA was extracted using the Zymo Direct-zol RNA kit with TRI reagent. mRNA was sequenced using the Hi-Seq Illumina platform with PolyA selection and a 2 × 150 bp configuration to yield >20 M pair-end reads per sample.

Fig. 1.

Fig 1 dummy alt text

Open in a new tab

A histology image from the experiment showing how the crypt depth and villus height were measured.

Statistical analysis and bioinformatics

Normality of each histological metric (villus height, crypt depth, and villus height:crypt depth ratio) was assessed with the Shapiro-Wilk test (α = 0.05), applied within each performance group (n = 13 high-performing, 13 low-performing). Normally distributed data were analysed with the lme package of R and multiple comparisons used the Sidak adjustment. A P < 0.05 significance level was used to establish significant differences. Other measures that departed from normality were analyzed with non-parametric tests: the Wilcoxon rank-sum (Mann-Whitney) test for the high- vs low-performing comparison and the Kruskal-Wallis test for the four-farm comparison. No transformations were applied to any variable.

Because villus height was measured in ∼100 µm increments, the coincident 900 µm medians across performance groups reflect measurement resolution rather than a lack of variation (Mann–Whitney P = 0.30).

For RNA-seq data, FastQC (version 0.11.9) was used to quality check raw reads. Data were cleaned using TrimGalore (RRID:SCR_011847) to remove poor quality portions of the reads and unwanted sequences such as poly-A tails or adapters (Martin, 2011). Pseudoalignment was conducted using Kallisto (Bray et al., 2016), whereby short sequences of bases (k-mers) were mapped to the reference transcripts and De Bruijn graphs were used to align overlapping sequences. Counts (the number of sequences assigned to each gene) were generated through alignment to the reference genome Gallus gallus version GRCg6a. Data normalization was conducted with DeSeq2 to account for technical variation due to differences in sequencing depth and mRNA composition (Love et al., 2014).

Differential gene expression between HP and LP farms were analysed using the DEseq2 package of R (Love et al., 2014). P-values from differential gene expression were adjusted using the Benjamini and Hochberg method (Benjamini and Hochberg, 1995) to correct for multiple testing and are presented as Padj. Statistical significance was considered at Padj < 0.05.

The enrichment of pathways from the Reactome database (Fabregat et al., 2018) was evaluated using the GSEA method (Subramanian et al., 2005), which calculates whether a set of genes from a pathway is overrepresented in the list of differentially expressed genes (DEGs), thereby assigning an enrichment score (Mootha et al., 2003). An average enrichment score was calculated for each performance category group. P-values were estimated for each pathway and then corrected for multiple testing (Padj) using the Benjamini and Hochberg method (Benjamini and Hochberg, 1995). Statistical significance was considered at Padj < 0.05.

In addition, a topology-based, Quantitative Pathway Activation method (QPA; Biofractal (Portugal), which implements strategies based on the pathway analysis methods evaluated by Ma et al. (2019), was used to derive biologically significant insights from DEGs (Klünemann et al., 2024). The method selects DEGs with a Padj <0.05, and then uses the expression levels of genes, their statistical significance, and their topological importance in the pathway to generate a pathway activation score. The activation score represents the number of standard deviations a given data point lies above (positive score) or below (negative score) the reference mean. One pathway activation score is calculated per sample. The pathway catalogue for the QPA method was a customized catalogue (Biofractal, Portugal) based on the Reactome database (Gillespie et al., 2022) with additions from Gene Ontology database (Ashburner et al., 2000) and published literature.

The integration of QPA data from jejunum and liver were conducted using DIABLO (Singh et al., 2019). Normalized data were fitted to performance category (LP vs HP farms) using sparse partial least square discriminant analysis (sPLS-DA). Sparsity in sPLS-DA was introduced by using Lasso penalty to minimize the number of selected features in each component and thus maximize their importance for discriminating samples according to performance category. The selected pathways were correlated using a similarity score that is analogous to a Pearson correlation coefficient (González et al., 2012).

Finally, a phenotypic association method (Biofractal, Portugal) was used to relate QPA data with phenotypic animal data. The method used linear estimation of associations to rank and quantify pathways associated with growth performance or intestinal histology using a regularized regression algorithm. Pathways that were associated with performance category (LP vs. HP farms), with the highest positive or negative association with phenotypic variables were identified and linear coefficients were estimated. These linear coefficients represent the size of the change in growth performance or histology that was associated with a change of one standard deviation in the activation of a particular pathway. To calculate the predicted effect size of each of these identified pathways on phenotypic variables, linear coefficients were multiplied by the activation score of samples. Phenotypic association is enabled by the per sample pathway activation scores that are generated using the QPA method. The models for differential gene expression and pathway analyses included performance category (LP vs HP) as a main effect and bird sex and RNA integrity as covariates.

Results

Animal performance and histology

Final flock performance is presented in Table 1. Overall, live body weights at 63 d were below the breed standard of 4,604 g for as-hatched flocks by 6.4 to 11.0 %, but part of that difference is explained by transportation loss. FCR were below the breed standard of 1.997 by 2.2 % to 4.8 %. When corrected by the body weight of the farm with the lowest BW, the lowest adjusted FCR was evident for farms 2 (1.868) and 3 (1.867), which were classified as HP farms for further analysis. Farms 1 and 4 had higher adjusted FCR (1.933 and 1.923, respectively) and were classified as LP farms. Total mortality varied between 4.32 and 6.54 % and was numerically greater for LP farms. Nonetheless, they were within a normal range of mortality for commercial broiler production at this age. The difference in adjusted FCR between LP and HP farms was 6 points of FCR. That difference was accompanied by a difference in final body weight of 136 g and a difference in total mortality of 1.5 percent (calculated differences in means of Low Performing and High Performing Farms, Table 1).

At the age of sampling (28 d), there was a difference of 2.8 g/d of ADG between LP and HP farms, reflected on a difference of 60 g of mean BW (Table 2).

At necropsy, only 3 birds from farm 3 (HP farm) and 1 bird from farm 4 (LP farm) presented E. maxima lesions in the jejunum (data not shown).

Crypt depth and the villus height:crypt depth ratio met the normality assumption in both groups (all P > 0.05) but villus height in the high-performing group departed from normality (W = 0.834, P = 0.018) (Supplementary Table 1).

Crypt depth was significantly greater for birds of the LP farms (381 µm; Table 3) compared to birds of the HP farms (270 µm), which was reflected in a lower villus height to crypt depth ratio for LP farms (2.69 versus 3.95, respectively). No difference in villi height was detected (Table 3).

Table 3.

Histology metrics of jejunal samples from broiler chickens from 4 farms, categorized by final performance based on FCR from hatch to 63 days of age.

Histology Metric Unit Farm 1 Farm 2 Farm 3 Farm 4 Probability1 High Performing2 Low Performing2 Probability1
Villus height µm 871 (850) 949 (900) 981 (900) 1,010 (1,000) 0.42 945 (900) 965 (900) 0.30
Crypt depth µm 392a 280b 262b 376a <0.001 270b 381a <0.001
Villus height:Crypt depth Ratio 2.53c 3.57a,b 4.29a 2.75c,b <0.001 3.95a 2.69b <0.001
Open in a new tab
a,b

Means with different superscripts differed at P < 0.05. Individual birds were considered experimental units.

1

Normally distributed data (crypt depth and villus height:crypt depth ratio) were analyzed with the lme package of R and multiple comparisons used the Sidak adjustment. For villus height, which was not normally distributed (Shapiro-Wilk, see Table S), values are given as the mixed-model estimated mean with the median in parentheses; the reported P-values are from the Kruskal-Wallis test (four-farm comparison) and the Wilcoxon rank-sum (Mann-Whitney) test (performance comparison), and no data transformations were applied. Categorical data were analyzed with chisq.test from R using the p.value simulation option.

2

Performance categories were established a-posteriori, based on adjusted FCR at 63 d of age.

Gene expression

Of the 64 samples collected from each tissue, 58 liver samples and 26 jejunum samples met the quality requirements for sequencing (RNA Integrity Score >6). This resulted in 31 liver samples for LP farms compared to 27 samples of HP farms, and for jejunum there were 13 samples from HP farms compared to 13 samples from LP farms.

A total of 361 and 1,148 DEGs were found in the jejunum and liver, respectively, of chickens from LP farms compared to chickens from HP farms (Fig. 2; Padj < 0.05). Among the upregulated genes in the liver of LP farms were several heat shock proteins and chaperones (Padj < 0.05; Fig. 3).

Fig. 2.

Fig 2 dummy alt text

Open in a new tab

Volcano plots showing; A. 361 differentially expressed genes in Low Performing birds compared to High Performing birds in jejunum tissue (Padj < 0.05).

B. 1,148 differentially expressed genes in Low Performing birds compared to High Performing birds in liver tissue (Padj < 0.05).

Fig. 3.

Fig 3 dummy alt text

Open in a new tab

Box and whisker plots showing the normalized counts of expression of genes related to heat shock proteins in the liver of High and Low Performing broilers. *Padj < 0.05. Central line indicates the median. N = 32 per group.

Pathway enrichment

The GSEA pathway enrichment algorithm, applied using the Reactome pathway database, showed many enriched pathways in the jejunum of birds from LP farms compared to HP farms. Positively enriched pathways (upregulated) in chickens from LP farms included those related to activation of the pre-replicative complex, cell cycle, haemostasis, immune system, metabolism of proteins, metabolism of RNA, signal transduction, and vesicle-mediated transport (Padj < 0.05; Fig. 4A). Conversely, negatively enriched (downregulated) pathways were related to digestion and absorption, drug ADME (absorption, distribution, metabolism, and excretion), metabolism, protein localization, sensory perception, and signal transduction (GTPases) (Padj <0.05; Fig. 4A).

Fig. 4.

Fig 4 dummy alt text

Open in a new tab

Pathway enrichment using Gene Set Enrichment Analysis showing enriched functions from the Reactome database in samples from Low Performing broilers compared to High Performing broilers (Padj < 0.05). A. Jejunum samples, B. Liver samples.

There were also many enriched pathways in the liver in relation to performance category. Positively enriched pathways in chickens from LP farms related to the cell cycle, cellular responses to stimuli including cellular response to heat stress and heat shock protein 90, developmental biology, extracellular matrix organization, haemostasis, immune system, interleukin receptor SHC signalling, muscle contraction, signal transduction, and vesicle mediated transport (Padj <0.05; Fig. 4B). Meanwhile, pathways relating to drug ADME, metabolism, metabolism of proteins, metabolism of RNA, organelle biogenesis and maintenance, and protein localisation were negatively enriched (Padj < 0.05; Fig. 4B).

Pathway activation and phenotypic association

The QPA data in the jejunum of chickens from LP farms showed activation of immune activation pathways (especially interferons and pro-inflammatory eicosanoids), gut integrity (especially cell replication), oxidative stress, and several amino acid metabolism related pathways compared to chickens from HP farms (Padj < 0.05; Fig. 5A). Conversely, pathways related to the digestion and absorption of fats, proteins, carbohydrates, and vitamins and minerals were inhibited in chickens from LP farms, alongside the metabolism of lipids, energy, and vitamins, and gut barrier function (tight junction proteins) (Padj <0.05; Fig. 5A).

Fig. 5.

Fig 5 dummy alt text

Open in a new tab

Quantitative pathway activation (QPA) showing activated (green) or inhibited (purple) functions from a custom Biofractal catalogue in samples from Low Performing broilers compared to High Performing broilers (Padj < 0.05). A. Jejunum samples, B. Liver samples.

The most important pathways that were related to the final body weight of birds included activation in hypoxia stress, WNT signalling, calcium signalling, host defence peptides, and inhibition in mitochondrial calcium ion transport and tight junction transmembrane proteins (Fig. 6A). Cumulatively, these pathways were predicted to be associated with a ∼70 g reduction in final body weight in the LP farms compared to birds from the HP farms (R2 0.82; Fig. 6A, Supplementary Table 2 Jejunum).

Fig. 6.

Fig 6 dummy alt text

Open in a new tab

Phenotypic association showing the predicted association between Quantitative Pathway Activation (QPA) and measured phenotypes of birds from Low and High Performing broilers (Padj < 0.05). A. Jejunum samples and final bird BW at 63 days of age. B. Jejunum samples and jejunum villi length to crypt depth ratio. C. Liver samples and final bird BW at 63 days of age. D. Liver samples and jejunum villi length to crypt depth ratio.

The most important pathways that were related to the jejunum villi length to crypt depth ratio were an inhibition of protein digestion enzymes, absorption of water-soluble vitamins, tight junction transmembrane proteins, micro-mineral absorption, and an activation in calcium signalling in chickens from LP compared to HP farms (Padj < 0.05; Fig. 6B). These pathway effects were predicted to be associated with lower villi length to crypt depth ratio in birds from LP farms compared to HP farms (R2 0.92; Fig. 6B).

In liver samples, pathways related to different aspects of immune function, liver integrity, oxidative stress, metabolic regulation, amino acid biosynthesis, sphingolipid metabolism, and lipid transport were activated in chickens from LP farms compared to HP farms (Padj < 0.05; Fig. 5B). Conversely, pathways related to energy, fatty acids and amino acid metabolism, JAK-STAT activation, cellular functions, and antioxidant systems were inhibited in birds from the Low Performing farms compared to the High performing farms (Padj < 0.05; Fig. 5B).

The phenotypic association between pathways in the liver and bird’s final body weight showed that activation in sphingolipid de novo biosynthesis, sensors of pathogen associated DNA, SUMOylation, hypoxia stress, oxidative stress induced senescence, heat stress, and inhibition of mitochondrial biogenesis in LP farms were predicted to be associated with a reduction in the final body weight of those birds by ∼40 g compared to birds from HP farms (R2 0.61; Fig. 6C, Supplementary Table 3 Liver).

Liver pathways were also associated with changes in the villi length to crypt depth ratio in the jejunum. Inhibition in glucagon signalling, mitochondrial fatty acid beta-oxidation, cysteine and homocysteine degradation, pyruvate metabolism, and antioxidant enzymatic systems in chickens from LP farms were predicted to be associated with lower villi length to crypt depth ratio in the jejunum compared to birds from HP farms (R2 0.58; Fig. 6D).

Integration of quantitative pathway activation data from different tissues

The integration of QPA data from jejunum and liver showed that a pathway related to gamma-aminobutyric acid (GABA) was activated in the liver of broilers from HP farms compared to LP farms (Padj < 0.05) and this was highly associated (R > 0.80) with nine pathways in the jejunum that were also activated in broilers from HP farms compared to LP farms (Padj < 0.05). The majority of activated pathways in the jejunum of broilers from HP farms, which were associated with activated GABA in the liver were related to fat metabolism, including PPAR alpha, vitamin A metabolism, mitochondrial fatty acid beta-oxidation, absorption of fat-soluble vitamins, transport of bile salts, portomicron assembly, and transport of fatty acids (Fig. 7).

Fig. 7.

Fig 7 dummy alt text

Open in a new tab

The integration of pathway activation data from jejunum (blue boxes) and liver (green boxes) tissue, showing positive (red lines) and negative (blue lines) correlations between functions in the two tissues (R > 0.80). Lines in the outer circle demonstrate whether the function was activated in the Low Performing birds (orange line on the outside) or activated in the High Performing birds (blue line on the outside) using Quantitative Pathway Activation.

In addition, several pathways related to amino acid metabolism in the liver were activated in broilers from HP farms (Padj < 0.05) and were associated with the inhibition of two cellular signalling pathways in the jejunum, SUMOylation and WNT signalling (R > 0.80; Fig. 7).

Discussion

To our knowledge, this is the first published proof of concept study on the ability of transcriptomics data to identify differences between LP and HP broiler chicken flocks in commercial production. Flocks of the same age within the same complex were selected to minimize sources of variation such as feed, which were equal amongst farms. Day 28 was selected as sampling age to provide a snapshot of the birds’ physiology immediately after the Eimeria cycles between 21 and 28 d, in an age of rapid growth, under the assumption that physiological differences at that age could significantly influence final performance.

There were no notable differences between birds from different performance categories at necropsy. However, the histology data suggested more subtle differences in jejunum morphology that could be related to performance. Birds from HP farms had lower crypt depth and greater villi length to crypt depth ratio, which suggested better intestinal health. The villi length to crypt depth ratio is a routine histological indicator of intestinal health (Jeurissen et al., 2002), and crypt depth has been associated with energy efficiency in commercial production (Metzler-Zebeli et al., 2018). A high villi length to crypt depth ratio is generally associated with a well differentiated intestinal mucosa and a high intestinal absorptive area (Jeurissen et al., 2002). The range of villi length to crypt depth ratios in the current study (2 to 4) was far lower than other observations in the range of 8 to 9 for similar age birds (Franco et al., 2006; Metzler-Zebeli et al., 2018; Rezar et al., 2024). The driver of this difference is the crypt depth, which was approximately 115 in those reports but 280 to 392 in the current study, indicating possible methodological differences.

A total of 1,148 and 361 DEGs were in the liver and jejunum, respectively, of LP farms compared to HP farms. The upregulated expression of heat shock proteins and chaperones in the liver of birds from LP farms indicated a possible stress challenge in those birds. However, it is difficult to interpret overall physiology from a small selection of genes. Two different pathway analysis methods, GSEA and QPA, were used to interpret the gene expression data in a more comprehensive way.

GSEA has been routinely used to understand systematic changes in gene function associated with biochemical pathways (Wu et al., 2021). GSEA assesses whether an a-priori defined set of genes shows statistically significant differences between two biological states and primarily determines whether the gene set is overrepresented among up- or down-regulated genes (Subramanian et al., 2005). Enrichment data provide a mean enrichment score per group, which limits the useability of these data for further analyses linking it to phenotypic variables within the population. Conversely, QPA provides a quantitative estimation of pathway activation or inhibition for an individual animal versus a reference population, considering the size or significance level of the changes in expression, and the importance of the genes in a pathway, which are not considered by enrichment methods (Mitrea et al., 2013).

The GSEA results in liver and jejunum samples showed some similarities. In both tissues, birds from LP farms had upregulated pathways related to immune function and cell cycle, but downregulated nutrient metabolism. Overall, this suggested a shift in the partitioning of resources in LP birds away from nutrient metabolism and toward immunity and cell turnover, which was also evident in the QPA analyses.

A large difference in the utility of the insights from QPA compared to GSEA was found in the ability of QPA to detect an inhibition in digestion and absorption pathways in the jejunum of chickens from LP farms, which aligns with the inhibited nutrient metabolism shown by both methods in this tissue. This difference on the insights is likely the result of a poorer annotation of digestion and absorption processes in the Reactome database that was remediated in the customized annotations used for the QPA method.

The increased cell replication in tissues noted by both GSEA and QPA methods is typically associated with regeneration and/or metabolic adaptation of the liver (Caldez et al., 2018), and increased rate of replacement of epithelial cells in the small intestine of chickens, which has been observed due to enteric issues such as an Eimeria challenge (Cloft et al., 2023; Fernando et al., 1973). These data suggest an increased level of damage and/or stress on the jejunum and liver tissues of birds from LP farms. The GSEA data in the liver identified an upregulation of genes associated with heat stress, in particular HSF-1 dependent transactivation, referring to the process by which heat shock factor 1 (HSF-1) activates the transcription of target genes in response to stress conditions, particularly heat shock. In chickens, HSF1 is activated by an intermediate level of heat stress, having a lower heat stress threshold compared to HSF3 (Tanabe et al., 1997). The attenuation of HSF1 occurs during continuous exposure to intermediate heat shock conditions or upon recovery from stress (Abravaya et al., 1991), which suggest that birds in the LP farms were exposed to a greater level of heat stress compared to HP birds prior to sampling. However, HSF1 has also been reported to activate gene expression in response to other types of stress, such as oxidative stress, inflammation, and infection (Shamovsky et al., 2008).

Although GSEA data provided no further insights on the source of cellular stress, QPA confirmed and specified cellular stress in both tissues. There was activation of pathways related to oxidative stress and hypoxia stress in both tissues, including the inhibition of antioxidant systems such as glutathione, antioxidant enzymes, and cytoprotection by HMOX1 in the liver. The phenotypic association analysis confirmed that these pathways were associated with changes in bird performance. In both tissues, activated hypoxia was associated with reduced final BW of LP birds, and in liver activation of oxidative stress and heat stress were further contributors to reduced final BW of these birds. Hypoxia in the liver is a known effect of inflammation and/or liver damage, whilst hypoxia in the intestine can lead to a degradation of epithelial barrier function (Ju et al., 2016). Activated immune cells, through their increased oxygen consumption and the oxygen consumption of reparative processes during intestinal inflammation, induce a state of hypoxia. Genes that are upregulated during hypoxia can have a direct effect on the function of tight junction proteins, thereby compromising the epithelial barrier (Manresa and Taylor, 2017). Supporting this hypothesis, QPA and phenotypic association showed an inhibition in barrier function and tight junction proteins in LP farms, which was predicted to be associated with reduced final BW in those birds.

The GSEA data suggested a negative relative nutritional balance of the LP birds by identifying the downregulation of pathways associated with energy metabolism, mitochondrial function, fatty acid metabolism, and amino acid metabolism in the liver. The QPA data showed a more clear pattern of inhibition of amino acid catabolism, and fatty acid metabolism, and provided further granularity, including activated synthesis of serine and methionine salvage, activated pathways related to plasma lipoproteins and HDL assembly, inhibited triglyceride biosynthesis and activated triglyceride catabolism, and inhibited mitochondrial beta-oxidation of fatty acids and gluconeogenesis but activated glycogen breakdown, glycolysis and utilization of ketone bodies. These metabolic changes point to a negative energy balance in the liver of LP birds compared to HP birds. For example, a reduction in amino acid utilisation for conservation, yet an increase in synthesis to support essential metabolism is commonly seen in the liver during fasting (Hou et al., 2020). Furthermore, the increased utilisation of ketones for energy has been observed in chickens fasted for 40 h (Lindholm, 2005) and may be associated with the observed inhibition in energy metabolism pathways. Ketone utilisation can decrease reliance on the energy metabolism pathways through pyruvate metabolism and the beta-oxidation of fatty acids for energy generation (Chandel et al., 2021). Interestingly, the inhibition of both pyruvate metabolism and mitochondrial beta-oxidation of fatty acids in the liver of LP broilers was associated with reduced villi length to crypt depth ratio, suggesting a link between overall nutritional balance and intestinal health.

Similarly, the inhibition of antioxidant systems and cysteine and homocysteine degradation in the liver were also associated with a reduced villus length to crypt depth ratio, indicating a link between depleted antioxidant capacity and worsened intestinal health in the LP farms. Indeed, a link between intestinal health, oxidative stress, and subsequent mitochondrial disfunction in the liver is recognised in the context of liver disease in humans (Das et al., 2023).

Alongside upregulated pathways on immune function, GSEA data showed that pathways relating to RHO GTPases were upregulated in the liver of birds from LP farms. Many pathogens, including viruses, have evolved capabilities to engage and subvert the actin cytoskeleton, in particular, the RHO family GTPase signalling system (Gouin et al., 2005), which has also been demonstrated in chickens (Liu et al., 2008; Rabiei et al., 2021). This effect could be suggestive of a pathogenic process, possibly viral. Furthermore, the QPA data highlighted activation of pathways related to interferon alpha and beta in both liver and jejunum, and additional activation of interferon gamma in jejunum. This immune response did not appear to be driven by an Eimeria maxima infection as evidenced by necropsy findings. The activation of interferons supports the hypothesis of a viral challenge in LP birds because members of the Type I IFN family have important antiviral roles (Goossens et al., 2013). Avian IFN-α inhibits the replication of many avian viruses such as AI, IBDV, IBV, MDV, and NDV, whereas avian IFN-β appears more potent than IFN-α for inducing IFN stimulated genes involved in signalling pathways (Qu et al., 2013).

The integration of QPA data from jejunum and liver showed interesting links between the two tissues, highlighting differences in systemic physiology between birds from LP and HP farms.

Birds from HP farms had activation in a pathway related to GABA in the liver. GABA is a four-carbon non-protein amino acid that acts as a primary inhibitory neurotransmitter in the central nervous system of animals, reducing the effects of stress. GABA receptors are present in non-neural tissues including the liver, pancreas and kidney allowing GABA to exert its antidiabetic, antioxidant, and immune modulating properties. In chickens, dietary GABA has been shown to minimize stress and improve gut function (Jeong et al., 2020). Activation of GABA pathways in the liver of HP birds suggests less stress in these birds compared to those from LP farms, and this was associated with many activated pathways related to fat absorption and transport in the jejunum. Research in mice has shown that the activation of stress in the liver or intestine can result in inhibited intestinal lipid absorption, as a result of specific transcription factors (Cheng et al., 2022). Therefore, the lower stress in HP birds indicated by activated GABA, may relate to increased lipid absorption in the intestine of these birds.

Other key connections highlighted in QPA data integration was the relationship between cell signalling pathways in the jejunum and amino acid metabolism in the liver. The cell signalling pathways SUMOylation and WNT signalling were activated in the jejunum of LP birds. SUMOylation is a post-translational protein modification that plays a major role in the WNT signaling pathway (Fan et al., 2022). WNT signaling is involved in cell differentiation, proliferation, and recovery, and plays a crucial role in the development and renewal of the intestinal epithelium (Mah et al., 2016). Indeed, the QPA and GSEA pathway analyses suggested an increased rate of replacement of jejunal epithelial cells, which has been observed due to enteric issues such as an Eimeria challenge (Cloft et al., 2023; Fernando et al., 1973). The data integration demonstrated that this increased epithelial turnover was associated with inhibited amino acid metabolism in the liver of LP birds, which could be a sign of reduced amino acid availability.

In conclusion, these data showed through both GSEA and QPA methods that birds from LP commercial farms had greater immune activation and relatively less active nutrient metabolism in the jejunum and liver compared to birds from HP commercial farms. These differences were reflected in jejunal histology, with birds from LP farms having greater crypt depth and a lower villi length to crypt depth ratio than birds from HP farms, and an inhibition of digestion and absorption processes in birds from LP farms identified by QPA. Inhibited protein digestion, and micro-mineral and water-soluble vitamin absorption processes were related to this change in jejunal morphology, and QPA data integration supported this finding, also indicating that higher jejunum turnover was associated with inhibited liver amino acid metabolism. Both the GSEA and QPA data suggested that LP birds were exposed to cellular stress. Whilst GSEA recorded this stress as pathways related to heat stress, the QPA method detected hypoxia and oxidative stress as sources of cellular stress that were directly associated with reduced final BW in LP birds. The increased granularity in insights from QPA comes from the functional and hierarchical nature of the custom pathway catalogue, the algorithms sensitivity to changes in pathways as a result of accounting for the expression level of genes, the statistical significance of genes, and their role in the topology of the pathway, and by generating a score per sample allowing association with phenotypic variables (e.g., performance, histology). Conversely, GSEA uses only the expression levels of genes below a certain P-value threshold, relies on non-tissue-specific pathway databases with higher complexity and a molecular biology-style annotation such as Reactome or Gene Ontology, and gives a mean enrichment score per sample, excluding the possibility of association with phenotypic data on an individual animal basis.

One limitation of the current work is the loss of a number of jejunum samples due to poor quality RNA, leading to a smaller samples size than planned. Whilst this does not affect the ability to compare between different methods (GSEA and QPA) it does mean that the biological inferences of jejunum samples and tissue integration should be treated as experimental hypotheses, and require validation with a larger sample number.

This proof-of-concept work led to hypotheses of how to improve the performance of LP farms in this study. For instance, the compromised oxidative status, intestinal barrier function, inflammation, and negative energy balance of LP flocks could be supported by dietary modification including alteration of energy or amino acid density, antioxidant supplementation, eubiotics, or phytogenics. Nonetheless, to be able to deliver robust insights that lead to an intervention and improved performance and health outcomes, validation of insights with phenotypic end points and interventions are necessary. In the future, it could even be possible to precisionly match the application of solutions to the problem, using complex omics data like gene expression in combination with advancing data science tools.

Ethics approval

The study was conducted under animal use protocols that complied with the Guide for the Care and Use of Agricultural Animals in Research and Teaching and were approved by Biofractal’s Animal Care and Use committee (protocol ANHb21-001).

Declaration of generative AI and AI-assisted technologies in the writing process

The authors did not use any artificial intelligence assisted technologies in the writing process.

Financial support statement

This work was supported by Biofractal Lda (Portugal).

CRediT authorship contribution statement

L.F. Romero: Writing – original draft, Methodology, Investigation, Funding acquisition, Conceptualization. C.E.C. Blue: Methodology, Investigation. D.A. Fenster: Methodology, Investigation. D. Ritter: Supervision, Resources, Methodology, Investigation, Conceptualization. L. Payling: Writing – original draft, Validation, Formal analysis, Data curation. K. Patel: Software, Formal analysis, Data curation. R.A. Dalloul: Writing – review & editing, Methodology, Conceptualization.

Disclosures

Luis Romero was serving as Director of Biofractal at the time of manuscript submission. The authors declare that aside from this professional affiliation, they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

The authors acknowledge and thank the integrated company that permitted the sampling and data analyses as well as the histopathology work on Dr. Frederic J. Hoerr in the project.

Footnotes

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.psj.2026.107522.

Appendix. Supplementary materials

Supplementary Table 1 shows the Shapiro-Wilk normality test results for jejunal histology metrics within each performance group.

mmc1.docx (15.1KB, docx)

Supplementary Table 2 Jejunum and Supplementary Table 3 Liver and contain Quantitative Pathway Activation scores, pathway regression coefficients, including standard errors, p-values, and confidence intervals and calculated predicted effect sizes on phenotypes for Low compared to High Performing Groups.

mmc2.xlsx (16.2KB, xlsx)
mmc3.xlsx (17.9KB, xlsx)

Data availability

None of the data were deposited in an official repository. All data are available upon request.

References

  1. Ashburner M., Ball C.A., Blake J.A., Botstein D., Butler H., Cherry J.M., Davis A.P., Dolinski K., Dwight S.S., Eppig J.T., Harris M.A., Hill D.P., Issel-Tarver L., Kasarskis A., Lewis S., Matese J.C., Richardson J.E., Ringwald M., Rubin G.M., Sherlock G. Gene ontology: tool for the unification of biology. Nat. Genet. 2000;25(1):25–29. doi: 10.1038/75556. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Aviagen Inc Ross 708 Performance guidelines. 2019. https://aviagen.com/na/brands/ross/products/ross-708 Accessed 10th January 2022 from.
  3. Abravaya K., Phillips B., Morimoto R.I. Attenuation of the heat shock response in HeLa cells is mediated by the release of bound heat shock transcription factor and is modulated by changes in growth and in heat shock temperatures. Genes Dev. 1991;5(11):2117–2127. doi: 10.1101/gad.5.11.2117. [DOI] [PubMed] [Google Scholar]
  4. Benjamini Y., Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J. R. Stat. Soc.: B (Methodol.) 1995;57:289–300. doi: 10.1111/j.2517-6161.1995.tb02031.x. [DOI] [Google Scholar]
  5. Bray N.L., Pimentel H., Melsted P., Pachter L. Near-optimal probabilistic RNA-seq quantification. Nat. Biotechnol. 2016;34(5):525–527. doi: 10.1038/nbt.3519. [DOI] [PubMed] [Google Scholar]
  6. Caldez M.J., Van Hul N., Koh H.W.L., Teo X.Q., Fan J.J., Tan P.Y., Dewhurst M.R., Too P.G., Talib S.Z.A., Chiang B.E., Stünkel W. Metabolic remodeling during liver regeneration. Dev. Cell. 2018;47(4):425–438.e5. doi: 10.1016/j.devcel.2018.09.020. e5. [DOI] [PubMed] [Google Scholar]
  7. Chandel N.S. Lipid metabolism. Cold Spring Harb. Perspect. Biol. 2021;13(9) doi: 10.1101/cshperspect.a040576. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cheng Y., Kang X.Z., Cheng T., Ye Z.W., Tipoe G.L., Yu C.H., Wong C.M., Liu B., Chan C.P., Jin D.Y. FACI is a novel CREB-H-induced protein that inhibits intestinal lipid absorption and reverses diet-induced obesity. Cell. Mol. Gastroenterol. Hepatol. 2022;13(5):1365–1391. doi: 10.1016/j.jcmgh.2022.01.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Cloft S.E., Miska K.B., Jenkins M., Proszkowiec-Weglarz M., Kahl S., Wong E.A. Temporal changes of genes associated with intestinal homeostasis in broiler chickens following a single infection with Eimeria acervulina. Poult. Sci. 2023;102(4) doi: 10.1016/j.psj.2023.102537. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Conesa A., Madrigal P., Tarazona S., Gomez-Cabrero D., Cervera A., McPherson A., Szcześniak M.W., Gaffney D.J., Elo L.L., Zhang X., Mortazavi A. A survey of best practices for RNA-seq data analysis. Genome Biol. 2016;17:13. doi: 10.1186/s13059-016-0881-8. https://link.springer.com/article/10.1186/s13059-016-0881-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Das A., De A.J., Mohanty T., Aich P. Role of oxidative stress, gut microbiota and derived metabolites in the etiology and progression of nonalcoholic fatty liver disease. Redox Exp. Med. 2023;2023(1) doi: 10.1530/REM-23-0016. [DOI] [Google Scholar]
  12. Fabregat A., Korninger F., Viteri G., Sidiropoulos K., Marin-Garcia P., Ping P., Hermjakob H. Reactome graph database: efficient access to complex pathway data. PLoS Comput. Biol. 2018;14(1) doi: 10.1371/journal.pcbi.1005968. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Fan L., Yang X., Zheng M., Yang X., Ning Y., Gao M., Zhang S. Regulation of SUMOylation targets associated with wnt/β-catenin pathway. Front. Oncol. 2022;12 doi: 10.3389/fonc.2022.943683. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Fernando M.A., McCraw B.M. Mucosal morphology and cellular renewal in the intestine of chickens following a single infection of Eimeria acervulina. J. Parasitol. 1973;59(3):493. doi: 10.2307/3278782. [DOI] [PubMed] [Google Scholar]
  15. Franco J.R.G., Murakami A., Natali M., Garcia E.R.M., Furlan A. Influence of delayed placement and dietary lysine levels on small intestine morphometrics and performance of broilers. Braz. J. Poult. Sci. 2006;8:233–241. doi: 10.1590/S1516-635X2006000400006. [DOI] [Google Scholar]
  16. Gillespie M., Jassal B., Stephan R., Milacic M., Rothfels K., Senff-Ribeiro A., Griss J., Sevilla C., Matthews L., Gong, Deng C. The reactome pathway knowledgebase 2022. Nucleic Acids Res. 2022;50(D1):D687–D692. doi: 10.1093/nar/gkab1028. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. González I., Cao K.A.L., Davis M.J., Déjean S. Visualising associations between paired ‘omics’ data sets. BioData Min. 2012;5(1):19. doi: 10.1186/1756-0381-5-19. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Goossens K.E., Ward A.C., Lowenthal J.W., Bean A.G. Chicken interferons, their receptors and interferon-stimulated genes. Dev. Comp. Immunol. 2013;41(3):370–376. doi: 10.1016/j.dci.2013.05.020. [DOI] [PubMed] [Google Scholar]
  19. Gouin E., Welch M.D., Cossart P. Actin-based motility of intracellular pathogens. Curr. Opin. Microbiol. 2005;8(1):35–45. doi: 10.1016/j.mib.2004.12.013. [DOI] [PubMed] [Google Scholar]
  20. Hou Y., Hu S., Li X., He W. In: Amino Acids in Nutrition and Health (Advances in Experimental Medicine and Biology, Vol. 1265) Wu G., Wu G., editors. Springer, Cham; 2020. Amino acid metabolism in the liver: nutritional and physiological significance.doi:10.1007/978-3-030-45328-2_2 [DOI] [PubMed] [Google Scholar]
  21. Huang W., Sherman B.T., Lempicki R.A. Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Res. 2009;37(1):1–13. doi: 10.1093/nar/gkn923. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Jeong S.B., Kim Y.B., Lee J.W., Kim D.H., Moon B.H., Chang H.H., Choi Y.H., Lee K.W. Role of dietary gamma-aminobutyric acid in broiler chickens raised under high stocking density. Anim. Nutr. 2020;6(3):293–304. doi: 10.1016/j.aninu.2020.03.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Jeurissen S.H., Lewis F., van der Klis J.D., Mroz Z., Rebel J.M., Ter Huurne A.A. Parameters and techniques to determine intestinal health of poultry as constituted by immunity, integrity, and functionality. Curr. Issues Intest. Microbiol. 2002;3(1):1–14. https://europepmc.org/article/med/12022808 [PubMed] [Google Scholar]
  24. Ju C., Colgan S.P., Eltzschig H.K. Hypoxia-inducible factors as molecular targets for liver diseases. J. Mol. Med. 2016;94(6):613–627. doi: 10.1007/s00109-016-1408-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Klünemann M., Romero L.F., Acman M., Milfort M.C., Fuller A.L., Rekaya R., Aggrey S.E., Payling L.M., Lemme A. Multitissue transcriptomics demonstrates the systemic physiology of methionine deficiency in broiler chickens. Animal. 2024;18(5) doi: 10.1016/j.animal.2024.101143. [DOI] [PubMed] [Google Scholar]
  26. Leeson S., Summers J.D. Nottingham University Press; Nottingham: 2005. Commercial Poultry Nutrition. [Google Scholar]
  27. Lindholm C. Intermittent Fasting in Chickens: Physiological Mechanisms and Welfare Implications for Broiler Breeders. Linköping University Electronic Press; Linköping, Sweden: 2005. ((Linköping Studies in Science and Technology Dissertation No. 2005)). [Google Scholar]
  28. Liu H.J., Lin P.Y., Wang L.R., Hsu H.Y., Liao M.H., Shih W.L. Activation of small GTPases RhoA and Rac1 is required for avian reovirus p10-induced syncytium formation. Mol. Cells. 2008;26(4):396–403. doi: 10.1016/S1016-8478(23)14014-3. [DOI] [PubMed] [Google Scholar]
  29. Love M.I., Huber W., Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15:550. doi: 10.1186/s13059-014-0550-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Ma J., Shojaie A., Michailidis G. A comparative study of topology-based pathway enrichment analysis methods. BMC Bioinform. 2019;20:546. doi: 10.1186/s12859-019-3146-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Mah A.T., Yan K.S., Kuo C.J. Wnt pathway regulation of intestinal stem cells. J. Physiol. 2016;594(17):4837–4847. doi: 10.1113/JP271754. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Manresa M.C., Taylor C.T. Hypoxia inducible factor (HIF) hydroxylases as regulators of intestinal epithelial barrier function. Cell. Mol. Gastroenterol. Hepatol. 2017;3(3):303–315. doi: 10.1016/j.jcmgh.2017.02.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Martin M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet. J. 2011;17(1):10. doi: 10.14806/ej.17.1.200. [DOI] [Google Scholar]
  34. Metzler-Zebeli B.U., Magowan E., Hollmann M., Ball M.E.E., Molnár A., Witter K., Ertl R., Hawken R.J., Lawlor P.G., O’Connell N.E., Aschenbach J. Differences in intestinal size, structure, and function contributing to feed efficiency in broiler chickens reared at geographically distant locations. Poult. Sci. 2018;97(2):578–591. doi: 10.3382/ps/pex332. [DOI] [PubMed] [Google Scholar]
  35. Mitrea C., Taghavi Z., Bokanizad B., Hanoudi S., Tagett R., Donato M., Voichiţa C., Drăghici S. Methods and approaches in the topology-based analysis of biological pathways. Front. Physiol. 2013;4:278. doi: 10.3389/fphys.2013.00278. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Mootha V.K., Lindgren C.M., Eriksson K.F., Subramanian A., Sihag S., Lehar J., Puigserver P., Carlsson E., Ridderstråle M., Laurila E., Houstis N. PGC-1α-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes. Na. Genet. 2003;34(3):267–273. doi: 10.1038/ng1180. [DOI] [PubMed] [Google Scholar]
  37. Qu H., Yang L., Meng S., Xu L., Bi Y., Jia X., Li J., Sun L., Liu W. The differential antiviral activities of chicken interferon α (ChIFN-α) and ChIFN-β are related to distinct interferon-stimulated gene expression. PLoS One. 2013;8(3) doi: 10.1371/journal.pone.0059307. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Rabiei M., Low W.Y., Ren Y., Cahyono M.I., Doan P.T.K., Dharmayanti I., Grande E.D., Hemmatzadeh F. Indicators of the molecular pathogenesis of virulent Newcastle disease virus in chickens revealed by transcriptomic profiling of spleen. Sci. Rep. 2021;11(1) doi: 10.1038/s41598-021-96929-w. https://www.nature.com/articles/s41598-021-96929-w [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Reimand J., Isserlin R., Voisin V., others Pathway enrichment analysis and visualization of omics data using g:profiler, GSEA, Cytoscape and EnrichmentMap. Nat. Protoc. 2019;14(482–517):482–517. doi: 10.1038/s41596-018-0103-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Rezar V., Pal M.P., Leskovec J., Levart A., Salobir J., Lavrenčič A., Vrecl M., Pirman T. Combined effects of cyclic heat stress, dietary induced oxidative stress and different levels of antioxidant on gut fermentation activity and mucosal morphology in broiler chickens. Agriculture. 2024;14(1):64. doi: 10.3390/agriculture14010064. [DOI] [Google Scholar]
  41. Shamovsky I., Nudler E. New insights into the mechanism of heat shock response activation. Cell. Mol. Life Sci. 2008;65:855–861. doi: 10.1007/s00018-008-7458-y. https://link.springer.com/article/10.1007/s00018-008-7458-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Singh A., Shannon C.P., Gautier B., Rohart F., Vacher M., Tebbutt S.J., Lê Cao K.A. DIABLO: an integrative approach for identifying key molecular drivers from multi-omics assays. Bioinformatics. 2019;35(17):3055–3062. doi: 10.1093/bioinformatics/bty1054. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Stanley D., Wu S.-B., Rodgers N., Swick R.A., Moore R.J. Subclinical necrotic enteritis: its aetiology and predisposing factors in commercial broiler production. World's. Poult. Sci. J. 2014;70(4):829–843. doi: 10.1017/S004393391400078X. [DOI] [Google Scholar]
  44. Subramanian A., Tamayo P., Mootha V.K., Mukherjee S., Ebert B.L., Gillette M.A., Mesirov J.P. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl. Acad. Sci. U. S. A. 2005;102(43):15545–15550. doi: 10.1073/pnas.050658010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Tanabe M., Nakai A., Kawazoe Y., Nagata K. Different thresholds in the responses of two heat shock transcription factors, HSF1 and HSF3. J. Biol. Chem. 1997;272(24):15389–15395. doi: 10.1074/jbc.272.24.15389. [DOI] [PubMed] [Google Scholar]
  46. Wu T., Hu E., Xu S., Chen M., Guo P., Dai Z., Feng T., Zhou L., Tang W., Zhan L.I., Fu X. clusterProfiler 4.0: a universal enrichment tool for interpreting omics data. Innovation. 2021;2(3) doi: 10.1016/j.xinn.2021.100141. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Table 1 shows the Shapiro-Wilk normality test results for jejunal histology metrics within each performance group.

mmc1.docx (15.1KB, docx)

Supplementary Table 2 Jejunum and Supplementary Table 3 Liver and contain Quantitative Pathway Activation scores, pathway regression coefficients, including standard errors, p-values, and confidence intervals and calculated predicted effect sizes on phenotypes for Low compared to High Performing Groups.

mmc2.xlsx (16.2KB, xlsx)
mmc3.xlsx (17.9KB, xlsx)

Data Availability Statement

None of the data were deposited in an official repository. All data are available upon request.

Articles from Poultry Science are provided here courtesy of Elsevier

RESOURCES