Limits...
Assessing Concordance of Drug-Induced Transcriptional Response in Rodent Liver and Cultured Hepatocytes.

Sutherland JJ, Jolly RA, Goldstein KM, Stevens JL - PLoS Comput. Biol. (2016)

Bottom Line: We examined the concordance of compound-induced transcriptional changes using data from several sources: rat liver and rat primary hepatocytes (RPH) from Drug Matrix (DM) and open TG-GATEs (TG), human primary hepatocytes (HPH) from TG, and mouse liver/HepG2 results from the Gene Expression Omnibus (GEO) repository.Co-expression networks performed better than genes or GSA when comparing treatment effects within rat liver and rat vs. mouse liver.We observe that the baseline state of untreated cultured cells relative to untreated rat liver shows striking similarity with toxicant-exposed cells in vivo, indicating that gross systems level perturbation in the underlying networks in culture may contribute to the low concordance.

View Article: PubMed Central - PubMed

Affiliation: Lilly Research Laboratories, Eli Lilly and Company, Indianapolis, Indiana, United States of America.

ABSTRACT
The effect of drugs, disease and other perturbations on mRNA levels are studied using gene expression microarrays or RNA-seq, with the goal of understanding molecular effects arising from the perturbation. Previous comparisons of reproducibility across laboratories have been limited in scale and focused on a single model. The use of model systems, such as cultured primary cells or cancer cell lines, assumes that mechanistic insights derived from the models would have been observed via in vivo studies. We examined the concordance of compound-induced transcriptional changes using data from several sources: rat liver and rat primary hepatocytes (RPH) from Drug Matrix (DM) and open TG-GATEs (TG), human primary hepatocytes (HPH) from TG, and mouse liver/HepG2 results from the Gene Expression Omnibus (GEO) repository. Gene expression changes for treatments were normalized to controls and analyzed with three methods: 1) gene level for 9071 high expression genes in rat liver, 2) gene set analysis (GSA) using canonical pathways and gene ontology sets, 3) weighted gene co-expression network analysis (WGCNA). Co-expression networks performed better than genes or GSA when comparing treatment effects within rat liver and rat vs. mouse liver. Genes and modules performed similarly at Connectivity Map-style analyses, where success at identifying similar treatments among a collection of reference profiles is the goal. Comparisons between rat liver and RPH, and those between RPH, HPH and HepG2 cells reveal lower concordance for all methods. We observe that the baseline state of untreated cultured cells relative to untreated rat liver shows striking similarity with toxicant-exposed cells in vivo, indicating that gross systems level perturbation in the underlying networks in culture may contribute to the low concordance.

Show MeSH

Related in: MedlinePlus

Concordance of drug-induced effects in rat and mouse liver.Pairs of experiments involving the same drug are compared via the Pearson R and percent overlap metrics, and average concordance is calculated between TG rat liver (reference set) and A) TG rat liver (within-source), B) DM rat liver (cross-source) and C) GEO mouse liver (cross-source and cross-species). Concordance is shown additively for 3 levels of constraint on dose and time differences: experiments at the same time point and doses within 5 fold (pink), time within 2 fold and doses within 10 fold (blue), no constraint on doses or time (green). In each case, experiments are compared to TG rat liver and the most concordant experiment selected within the level of constraint. Only experiments that have a match to TG rat liver at the most constrained level are considered at the other levels. Because concordance can only improve upon removing dose and time constraints, results are shown additively with the total bar height denoting the concordance achieved with no constraints. Numeric values represent the probability that the given level of concordance exceeds that of random pairs involving different drugs for the most stringent level considered (pink bars). Experiment pairs are separated into 3 ranges based on the level of transcriptional activity for the least perturbing of two treatments (lowest quartile have avg. abs. EG ≤ 0.28, highest quartile have avg. abs. EG > 0.46; thresholds selected using 3528 TG rat liver experiments).
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4814051&req=5

pcbi.1004847.g002: Concordance of drug-induced effects in rat and mouse liver.Pairs of experiments involving the same drug are compared via the Pearson R and percent overlap metrics, and average concordance is calculated between TG rat liver (reference set) and A) TG rat liver (within-source), B) DM rat liver (cross-source) and C) GEO mouse liver (cross-source and cross-species). Concordance is shown additively for 3 levels of constraint on dose and time differences: experiments at the same time point and doses within 5 fold (pink), time within 2 fold and doses within 10 fold (blue), no constraint on doses or time (green). In each case, experiments are compared to TG rat liver and the most concordant experiment selected within the level of constraint. Only experiments that have a match to TG rat liver at the most constrained level are considered at the other levels. Because concordance can only improve upon removing dose and time constraints, results are shown additively with the total bar height denoting the concordance achieved with no constraints. Numeric values represent the probability that the given level of concordance exceeds that of random pairs involving different drugs for the most stringent level considered (pink bars). Experiment pairs are separated into 3 ranges based on the level of transcriptional activity for the least perturbing of two treatments (lowest quartile have avg. abs. EG ≤ 0.28, highest quartile have avg. abs. EG > 0.46; thresholds selected using 3528 TG rat liver experiments).

Mentions: Concordance of experiment pairs is tabulated across quartiles of transcriptional activity for the least-perturbing of two treatments in the pair, given the importance of this variable in explaining intra-source concordance (Fig 2, S2 Dataset). This can be seen from comparisons within TG rat liver, where all methods yield higher concordance with increasing levels of transcriptional activity (Fig 2A). Improvement in concordance when relaxing constraints on dose and time differences (pink -> blue -> green bars) is minimal within-source, indicating that the most similar experiments by gene expression tend to be the most similar experiments by dose/time. All methods have high self-identification probability.


Assessing Concordance of Drug-Induced Transcriptional Response in Rodent Liver and Cultured Hepatocytes.

Sutherland JJ, Jolly RA, Goldstein KM, Stevens JL - PLoS Comput. Biol. (2016)

Concordance of drug-induced effects in rat and mouse liver.Pairs of experiments involving the same drug are compared via the Pearson R and percent overlap metrics, and average concordance is calculated between TG rat liver (reference set) and A) TG rat liver (within-source), B) DM rat liver (cross-source) and C) GEO mouse liver (cross-source and cross-species). Concordance is shown additively for 3 levels of constraint on dose and time differences: experiments at the same time point and doses within 5 fold (pink), time within 2 fold and doses within 10 fold (blue), no constraint on doses or time (green). In each case, experiments are compared to TG rat liver and the most concordant experiment selected within the level of constraint. Only experiments that have a match to TG rat liver at the most constrained level are considered at the other levels. Because concordance can only improve upon removing dose and time constraints, results are shown additively with the total bar height denoting the concordance achieved with no constraints. Numeric values represent the probability that the given level of concordance exceeds that of random pairs involving different drugs for the most stringent level considered (pink bars). Experiment pairs are separated into 3 ranges based on the level of transcriptional activity for the least perturbing of two treatments (lowest quartile have avg. abs. EG ≤ 0.28, highest quartile have avg. abs. EG > 0.46; thresholds selected using 3528 TG rat liver experiments).
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4814051&req=5

pcbi.1004847.g002: Concordance of drug-induced effects in rat and mouse liver.Pairs of experiments involving the same drug are compared via the Pearson R and percent overlap metrics, and average concordance is calculated between TG rat liver (reference set) and A) TG rat liver (within-source), B) DM rat liver (cross-source) and C) GEO mouse liver (cross-source and cross-species). Concordance is shown additively for 3 levels of constraint on dose and time differences: experiments at the same time point and doses within 5 fold (pink), time within 2 fold and doses within 10 fold (blue), no constraint on doses or time (green). In each case, experiments are compared to TG rat liver and the most concordant experiment selected within the level of constraint. Only experiments that have a match to TG rat liver at the most constrained level are considered at the other levels. Because concordance can only improve upon removing dose and time constraints, results are shown additively with the total bar height denoting the concordance achieved with no constraints. Numeric values represent the probability that the given level of concordance exceeds that of random pairs involving different drugs for the most stringent level considered (pink bars). Experiment pairs are separated into 3 ranges based on the level of transcriptional activity for the least perturbing of two treatments (lowest quartile have avg. abs. EG ≤ 0.28, highest quartile have avg. abs. EG > 0.46; thresholds selected using 3528 TG rat liver experiments).
Mentions: Concordance of experiment pairs is tabulated across quartiles of transcriptional activity for the least-perturbing of two treatments in the pair, given the importance of this variable in explaining intra-source concordance (Fig 2, S2 Dataset). This can be seen from comparisons within TG rat liver, where all methods yield higher concordance with increasing levels of transcriptional activity (Fig 2A). Improvement in concordance when relaxing constraints on dose and time differences (pink -> blue -> green bars) is minimal within-source, indicating that the most similar experiments by gene expression tend to be the most similar experiments by dose/time. All methods have high self-identification probability.

Bottom Line: We examined the concordance of compound-induced transcriptional changes using data from several sources: rat liver and rat primary hepatocytes (RPH) from Drug Matrix (DM) and open TG-GATEs (TG), human primary hepatocytes (HPH) from TG, and mouse liver/HepG2 results from the Gene Expression Omnibus (GEO) repository.Co-expression networks performed better than genes or GSA when comparing treatment effects within rat liver and rat vs. mouse liver.We observe that the baseline state of untreated cultured cells relative to untreated rat liver shows striking similarity with toxicant-exposed cells in vivo, indicating that gross systems level perturbation in the underlying networks in culture may contribute to the low concordance.

View Article: PubMed Central - PubMed

Affiliation: Lilly Research Laboratories, Eli Lilly and Company, Indianapolis, Indiana, United States of America.

ABSTRACT
The effect of drugs, disease and other perturbations on mRNA levels are studied using gene expression microarrays or RNA-seq, with the goal of understanding molecular effects arising from the perturbation. Previous comparisons of reproducibility across laboratories have been limited in scale and focused on a single model. The use of model systems, such as cultured primary cells or cancer cell lines, assumes that mechanistic insights derived from the models would have been observed via in vivo studies. We examined the concordance of compound-induced transcriptional changes using data from several sources: rat liver and rat primary hepatocytes (RPH) from Drug Matrix (DM) and open TG-GATEs (TG), human primary hepatocytes (HPH) from TG, and mouse liver/HepG2 results from the Gene Expression Omnibus (GEO) repository. Gene expression changes for treatments were normalized to controls and analyzed with three methods: 1) gene level for 9071 high expression genes in rat liver, 2) gene set analysis (GSA) using canonical pathways and gene ontology sets, 3) weighted gene co-expression network analysis (WGCNA). Co-expression networks performed better than genes or GSA when comparing treatment effects within rat liver and rat vs. mouse liver. Genes and modules performed similarly at Connectivity Map-style analyses, where success at identifying similar treatments among a collection of reference profiles is the goal. Comparisons between rat liver and RPH, and those between RPH, HPH and HepG2 cells reveal lower concordance for all methods. We observe that the baseline state of untreated cultured cells relative to untreated rat liver shows striking similarity with toxicant-exposed cells in vivo, indicating that gross systems level perturbation in the underlying networks in culture may contribute to the low concordance.

Show MeSH
Related in: MedlinePlus