Skip to main content

Human genomics. The human transcriptome across tissues and individuals.

Melé M, Ferreira PG, Reverter F, DeLuca DS, Monlong J, Sammeth M, Young TR, Goldmann JM, Pervouchine DD, Sullivan TJ, Johnson R, Segrè AV, Djebali S, Niarchou A, Wright FA, Lappalainen T, Calvo M, Getz G, Dermitzakis ET, Ardlie KG, Guigó R

Publication: Science (2015) • Vol 348 • Page 660-5PMID: 25954002View on PubMed

Purpose

The purpose of this paper was to determine the effects of patient age, sex and race, tissue type, ischemic time and analysis platform on RNA expression and splicing in postmortem PAXgene-preserved specimens and to compare data with publically available data from healthy living donors.

Conclusion

Expression of protein coding RNAs (PCGs) and long noncoding RNAs (lncRNAs) clustered by tissue type, with postmortem tissues clustering with those of the same type from living surgical donors. While 47% of the variability in expression was attributable to tissue type, only 4% was attributed to individual variability. Individual PCGs and lncRNAs with sex, race, age and tissue-specific expression patterns were identified. 87,000 novel splice junctions were identified and were found to have tissue-specific expression, the majority of which were located in the cerebellum. Differential splicing had more unexplained variability (not attributable to individual or tissue type) than differential expression. Importantly, the ischemic time differed between specimens that clustered based on exon exclusion/inclusion. Expression levels were strongly correlated between RNA-seq and Affymetrix microarray data (average correlation coefficient=0.83).

Studies (1)

Purpose

The purpose of this study was to determine the effects of patient age, sex and race, tissue type, ischemic time and analysis platform on RNA expression and splicing in postmortem PAXgene-preserved specimens and to compare data with publically available data from healthy living donors. Sequencing data was generated for 1641 frozen PAXgene-preserved specimens from 175 postmortem individuals. Specimens included 29 solid tissues, 11 brain regions, blood, cultured fibroblasts and Epstein-Barr virus–transformed lymphocytes, with analysis priority given to adipose, tibial artery, heart, lung, skeletal muscle, tibial nerve, skin, and thyroid specimens. Only specimens with a RNA integrity number (RIN) >6 were included. An average of 80 million paired-end mapped reads were generated for each specimen, and results were compared to microarray expression profiles for 609 normal specimens in the Gene Expression Omnibus (GEO) database.

Summary of Findings

PCGs tended to be expressed in all 9 priority tissues, but lncRNAs tended to have tissue-specific expression patterns or weren't expressed in the priority tissues. Of the fewer than 200 genes expressed exclusively in one tissue, 95% were exclusive to testis. Expression clustered by tissue type, even when postmortem tissues were compared with specimens from living donors, with a primary separation observed between blood and solid tissues. The brain was the most distinct of the solid tissues, but with the exception of cerebellum, brain regions clustered together. Overall, 47% of the variation in expression was between tissue types and 4% was between individuals. Expression patterns of PCGs and lncRNAs were dependent on sex, race or age and were tissue-specific sex based (mostly in breast) and tissue-specific ethnicity based (predominantly in skin). 88 sex-dependent coexpression networks were found. Most splice junctions identified had tissue-specific expression. The brain accounted for 40% of these variants overall, and the brain-specific expression was particularly pronounced for micro-exons (15-60 bp) and often localized to the cerebellum. Importantly, splicing had more unexplained variability (not attributable to individual or tissue type) than expression levels, and this was especially pronounced for genes involved in translation or protein synthesis. The authors propose that ischemic time may be one factor contributing to the variation in splicing as when specimens clustered by exon inclusion within a tissue type, the clusters had different average ischemic times. Expression, as determined by RNA-seq and Affymetrix microarray data, was strongly correlated (average correlation coefficient=0.83).

  • Biospecimens

    TissueSpinal Cord
    TissueAdipose
    TissueAdrenal Gland
    TissueHeart
    TissueLung
    TissueMuscle (Skeletal)
    TissueNerve
    TissueSkin
    TissueThyroid Gland
    TissueTestis
    TissueLiver
    TissuePituitary Gland
    TissueBrain
    TissuePancreas
    TissueKidney
    TissueStomach
    TissueColorectal
    TissueVagina
    TissueOvary
    TissueOther
    TissueProstate
    TissueUterus
    TissueEsophagus
    TissueFallopian Tube
    FluidBlood
    TissueBlood Vessel
  • Preservation Types

    PAXgene
  • Diagnoses

    Autopsy
    Normal
    Not specified
  • Platforms

    RNANext generation sequencing
    RNADNA microarray
  • Pre-analytical Factors

    ClassificationPre-Analytical FactorValue(s)
    PreaquisitionPatient age
    20-70 years
    PreaquisitionPatient gender
    Female
    Male
    PreaquisitionPatient race
    African-American ancestry
    European ancestry
    PreaquisitionPostmortem interval
    0 h
    Tissue dependent average of 173-869 min
    Biospecimen AcquisitionBiospecimen location
    Adipose
    Amygdala
    Anterior cingulate cortex
    Blood
    Brain
    Caudate
    Cerebellum
    Cerebral hemishphere
    Colon
    Cortex
    Esophagus
    Fallopian Tube
    Frontal cortex
    Heart
    Hippocampus
    Hypothalamus
    Kidney
    Liver
    Lung
    Ovary
    Pancreas
    Pituitary
    Prostate
    Putamen
    Skeletal muscle
    Skin
    Spinal cord
    Stomach
    Substantia nigra
    Testis
    Thyroid
    Tibial artery
    Tibial nerve
    Uterus
    Vagina
    Biospecimen AcquisitionCold ischemia time
    0-1400 minutes
    300-1100 minutes
    Next generation sequencing SpecificTechnology platform
    Affymetrix microarray