| Download datasets and supplementary data files |
|---|
Summary
Our understanding of the evolution, formation, and pathological disruption of human brain circuits is impeded by a lack of comprehensive data on the developing brain transcriptome. Thus, we have undertaken whole-genome, exon-level expression analysis of thirteen regions from left and right sides of the mid-fetal human brain, finding 76% of genes to be expressed, and 44% of these to be differentially regulated. These data reveal a large number of specific gene expression and alternative splicing patterns, as well as co-expression networks, associated with distinct regions and neurodevelopmental processes. Of particular relevance to cognitive specializations, we have characterized the transcriptional landscapes of prefrontal cortex and perisylvian speech and language areas, which exhibit a population-level global expression symmetry. Finally, we show that differentially expressed genes are more frequently associated with human-specific evolution of putative cis-regulatory elements. Altogether, these data provide a wealth of novel biological insights into the complex transcriptional and molecular underpinnings of human brain development and evolution.
Experiment design
Human Brain Specimens and Tissue Processing This study was carried out using post-mortem human brain specimens collected from the Human Fetal Tissue Repository at the Albert Einstein College of Medicine (AECOM). Dissected tissue was fresh-frozen in Trizol for RNA and DNA extraction, with a post-mortem interval of less than 1 hour. Remaining tissue was fixed and frozen, and sections were analyzed for neuropathological or developmental defects. Details of specimens, tissue processing, microdissection, and neuropathological assessment are given in the Supplemental Experimental Procedures and Table S1. These studies were approved by the Human Investigation Committees of AECOM and Yale University.
For initial analysis, all neocortex samples were grouped together. In subsequent analyses, neocortex areas were compared with each other. In most analyses reported, left and right side samples were treated as additional biological replicates.
About cases
RNA was isolated from 13 brain regions, from both hemispheres, of four late mid-fetal human brains, with a total PMI of less than one hour, and hybridized to Affymetrix Human Exon 1.0 ST arrays. Affymetrix CEL files were imported into Partek GS using Robust Multichip Average (RMA) background correction, quantile normalization, and GC content correction. The normalized data were then converted to log-ratios, relative to arrays hybridized with RNA pooled from all regions of the same brain. Signal log-ratios are displayed here as green for negative (underexpression) and red for positive (overexpression).
Table 1 | Periods of human development and adulthood as defined in this study.
| Period | Description | Age |
| 1 | Embryonic | 4≤ Age <8 Postconceptual weeks (PCW) |
| 2 | Early fetal | 8≤ Age <10 PCW |
| 3 | Early fetal | 10≤ Age <13 PCW |
| 4 | Early midfetal | 13≤ Age <16 PCW |
| 5 | Early midfetal | 16≤ Age <19 PCW |
| 6 | Late midfetal | 19≤ Age <24 PCW |
| 7 | Late fetal | 24≤ Age <38 PCW |
| 8 | Neonatal and early infancy | Birth≤ Age <6 Postnatal months (M) |
| 9 | Late infancy | 6 M≤ Age <12 M |
| 10 | Early childhood | 1≤ Age <6 Postnatal years (Y) |
| 11 | Middle and late childhood | 6≤ Age <12 Y |
| 12 | Adolescence | 12≤ Age <20 Y |
| 13 | Young adulthood | 20≤ Age <40 Y |
| 14 | Middle adulthood | 40≤ Age <60 Y |
| 15 | Late adulthood | 60 Y ≤ Age |
About tissue
Tissue was microdissected from 13 regions, including 9 distinct neocortical areas, from both left and right sides of four late second trimester human brain specimens. Gene- and exon-level differential expression analyses were performed by mixed model, nested analysis of variance using the XRAY software from Biotique Systems. Further details available in Johnson, Kawasawa, et al., "Functional and Evolutionary Insights into Human Brain Development through Global Transcriptome Analysis" Neuron, Volume 62, Issue 4, 2009.
Ethnicity codes are as follows: AA = African American, A = , A/E = X, As,= , H = , E = , CC = and n/a= unknown
This track displays exon microarray expression data from the late mid-fetal human brain, generated by the Sestan Lab at Yale University. The data represent 13 brain regions, including nine areas of neocortex, and both hemispheres. By default, arrays are grouped by the median for each brain region, including each neocortical area. Alternatively, neocortex areas can be grouped together; arrays can be grouped by mean; or all 95 arrays can be shown individually.
Table 2 | Ontology and nomenclature of analyzed brain regions and neocortical areas.
| Period 1 | Period 2 | Period 3 - 15 |
| FC: Frontal cerebral wall | FC | OFC: Orbital prefrontal cortex |
| DFC: Dorsolateral prefrontal cortex | ||
| VFC: Ventrolateral prefrontal cortex | ||
| MFC: Medial prefrontal cortex | ||
| M1C: Primary motor (M1) cortex | ||
| PC: Parietal cerebral wall | PC | S1C: Primary somatosensory (S1) cortex |
| IPC: Posterior inferior parietal cortex | ||
| TC: Temporal cerebral wall | TC |  A1C: Primary auditory (A1) cortex |
| STC: Posterior superior temporal cortex | ||
| ITC: Inferior temporal cortex | ||
| OC: Occipital cerebral wall | OC | V1C: Primary visual (V1) cortex |
| HIP: Hippocampal anlage | HIP | HIP: Hippocampus |
| AMY: Amygdala | ||
| VF: Ventral forebrain | CGE: Caudal ganglionic eminence | STR: Striatum |
| LGE: Lateral ganglionic eminence | ||
| MGE: Medial ganglionic eminence | ||
| DIE: Diencephalon | DTH: Dorsal thalamus | MD: Mediodorsal nucleus of thalamus |
| URL: Upper (rostral) rhombic lip | URL | CBC: Cerebellar cortex |
About data processing
Normalized method: Quantile normalization. Partek Genomics Suite version 6.5 (Partek Incorporated, St. Louis, MO, USA) was used to normalize raw exon array data and to summarize expression of the probe set and transcript cluster. Affymetrix CEL files that passed QC analyses were imported into Partek Genomics Suite using the default Partek settings: RMA background correction114, quantile normalization, mean probe set summarization, and log2-transformation. Only high-quality core probe sets, as defined by Affymetrix, were included. 105,271 core probes (within 62,448 probe sets out of 230,918 core probe sets) contained SNPs defined in the probe group file HuEx-1_0-st-v2.r2-SNPs-Excluded.pgf provided by Affymetrix, which is based on the dbSNP database (version 129, April 2008) and SNPinprobe_1.0 database. we removed SNP-containing probe sets during the normalization step in the Partek program to be control for SNP-related confounding effects. The median of all individual probe sets of one transcript cluster was used as the estimate of gene expression values.
Contributors
Citation
- Johnson MB, Kawasawa YI, Mason CE, Krsnik Z et al. Functional and evolutionary insights into human brain development through global transcriptome analysis. Neuron 2009 May 28;62(4):494-509. PMID: 19477152
Notes
The raw microarray data are available via the NCBI Gene Expression Omnibus. Data in GeneNetwork as of July 2015 are only gene-level data. We could also enter exon-level data on request.