Analysis of gene regulation and cell fate from single-cell gene expression profiles in C. elegans

 

Figures

Click on the images for enlarged versions

Figure 1. Generation of cell lineage expression profiles.  A. Three dimensional image of a worm expressing mCherry from a promoter of interest, GFP in muscle cells from the myo-3 promoter, and stained with DAPI.  B.  The confocal image is computationally straightened and set to a standard three dimensional size.  C. Nuclei stained with DAPI are automatically identified, and then the identity of each nucleus is annotated with the help of WANO, a worm annotation tool (Peng et al., 2009). Nuclei were labeled by pseudo-colors for visualization.  D. Expression of the mCherry reporter in each of the 363 identified nuclei is calculated and displayed as a heat map.

Figure 2. A profile of gene expression at single cell resolution.  Shown is the mCherry expression level for 93 reporter genes in 363 cells (out of 558 total, 64%) in newly-hatched L1 larvae.  To make full use of the dynamic range in the heat map and to filter out background noise, we adjusted the gene expression level by calculating (gene expression level + 500)/500.  The scale bar shows log2(adjusted gene expression level).  Genes were clustered according to their expression profile.  Cells were manually arranged according to their tissue types.  Within each tissue type, cells were arranged according to their anterior-posterior order.  Bold indicates genes that have been previously analyzed for patterns of gene expression (Supplemental Table  1).  Red indicates genes that are expressed mainly in the body wall muscle cells.  Underline indicates cell lineage specific genes that are discussed in text.  Data for single cell gene expression including names of cells and expression in each individual worm can be found in Supplemental Tables 2-4.   b.w.m.: body wall muscle; blast: blast cells; epi.: epithelial cells; hyp.: hypodermal cells; int.; intestine cells; neu.: neurons; ph. m.: pharyngeal muscle.
Figure 3. Cell lineage dependent gene expression among nuclei in the hyp7 syncytium.  A. Genes are differentially expressed between AB-derived and C-derived nuclei (p< 10-5, t-test). Color indicates level of expression.  Gene expression levels were normalized for each gene so that the minimal and maximal expression values are 0 and 1 for each gene.  Expression levels in these nuclei and P-values for all genes are in Supplemental Table 5A. B.  Different transcriptional control of AB- versus C-derived nuclei in the hyp7 syncytium.  Shown are the tail area of L1 stage worms expressing a col-93:mCherry reporter (expressed in hypodermal cells), ajm-1:GFP, (a hypodermal cell boundary marker) and stained with DAPI.  The top and middle images show C08B11.3(RNAi) animals in which the two AB-derived nuclei do not express col-93:mCherry.  In most cases, the AB-derived nuclei have not fused with the syncytium (top) but sometimes they fuse (middle).  The bottom image shows a control with normal expression of col-93:mCherry in the hyp7 nuclei and cell fusion .   arrow head: AB-derived hyp7 nuclei; arrow: C-derived hyp7 nucleus.
Figure 4. Clustering diagram of 363 cells according to their gene expression profiles. The terrain map of nuclei was generated by Genesis {Sturn, 2002 #88}. Colors indicate different tissue types. Distance between cells in the x-y plane indicate levels of molecular similarity, such that cells with similar gene expression patterns are placed close to each other and cells with different patterns are placed far apart.
Figure 5. Molecular differentiation map for the C. elegans cell lineage.  A.  The developmental state of a cell is based on whether or not the cell is committed to express a gene, summed over all 93 of the reporter genes.  Each cell in each bifurcation in the tree (corresponding to one or more cell divisions) is compared to its sister cell to determine whether they have similar or different gene expression states.  Line thickness indicates degree of dissimilarity between these cells. The modified cell lineage is displayed. Dotted lines show the portions of the complete cell lineage that were unscored. The solid lines represent the modified lineage used for the analysis. Asterisks denote 16 new terminal cell divisions that are asymmetric. B.  Expanded view of the P1 lineage, showing the fates of cells.  Specific cells discussed in the text are shown.  Colored bars represent tissue types of terminal cell nodes in the cell lineage tree.
Fig. 6
Figure 6. Developmental clones and sublineages in the C. elegans lineage.  (A) Cells are aligned according to their lineage along the x- and y-axes. The similarity of a pair of cells is a measurement of the differences in the expression profile of the 93 reporter genes (ap defined in the Experimental Procedures).  The lineage of the first several cell divisions is shown.  Red boxes represent clones of cells of similar cell fate.  Blue boxes represent developmental sublineages.  Omitted cells are excluded from the map.  (B) AB.pl and AB.pr share a common sublineage.  (C) Developmental clones and sublineages from C.a and C.p.
Figure S1. Consistency of gene expression measurements. (A) Correlations of gene expression between different individual worms from the same transgenic line.  For every transgenic line, the Pearson correlation coefficient (R) for gene expression between all individual worms was calculated and averaged.  (B) Comparing expression correlations between worms from different transgenic lines derived from the same mCherry reporter construct with correlation among worms from the same transgenic line.  For 12 mCherry reporters, multiple integrated transgenic lines were generated and used for single-cell gene expression analysis.  The Pearson correlation coefficient for single cell gene expression between different worms was calculated.  The y-axis shows the mean of the Pearson correlation coefficients between worms of same transgenic line, and the x-axis shows the mean of the Pearson correlation coefficients between worms of different transgenic lines.  Error bars represent 95% confidential interval.
Figure S2. Heterogeneous gene expression patterns among body wall muscle cells.  A.  Different expression patterns among body wall muscle cells based on their lineage.  Columns represent different muscle cells, arranged by their lineage ancestry. Rows represent genes that are differentially expressed between muscles derived from different MS and D blastomeres (p< 10-5, t-test).   Gene expression levels were normalized for each gene so that the minimal and maximal expression values are 0 and 1 for each gene.  Expression levels in these nuclei and P-values for all genes are in Supplemental Table 5B.  B.  An anterior-posterior gradient of gene expression in the body wall muscles.  Rows represent different muscle cells, arranged from anterior to posterior.  Rows represent genes that are differentially expressed in the anterior-posterior axis (p< 10-5, linear regression), excluding lineage-specific genes shown in Supplemental Figure 5A.  Color indicates level of expression.  Expression levels in these nuclei and P-values for all genes are in Supplemental Table 5C.
test
Figure S3.Expression of C08B11.3 in hypodermis 7 nuclei.  A.  C08B11.3 maintains expression in AB-derived hyp 7 nuclei at least until the end of L1 stage.  Specific data for the table shown are shown in Supplemental Table 6.  B.  Expression of C08B11.3 re-appears following photobleaching.  The left figure shows expression C08B11.3:mCherry in hyp7 in a newly hatched L1 worm.  The worms were photobleached to remove fluorescence from pre-existing mCherry protein (middle figure) and mCherry expression re-appeared 15 hours later (right figure) showing that mCherry fluorescence in the L1 involves new protein synthesis and is not solely due to residual protein from the embryonic AB lineage. 
Sup. Fig. 4 Figure S4.  Gene expression commitment in the embryo.  For each gene, the gene commitment algorithm is used to predict commitment to express the gene in the embryonic lineage based on observed expression in the L1 larvae.  Shown is the embryonic lineage. Dotted lines show the portions of the complete cell lineage that were unscored. The solid lines represent the modified lineage used for the analysis. Red indicates commitment to express the gene.  O indicates cells in which expression has been previously observed and X indicates cells in which expression has previously been shown to shut (Ardizzi and Epstein, 1987; Hallam et al., 2000; Horner et al., 1998; Kalb et al., 1998; Murray et al., 2008).  Dashes indicate cells that were not scored.
Sup. Fig. 5

Figure S5. Gene expression commitment in the embryo for each of the 93 genes. 

SFig 6B Figure S6. Commitment Algorithm: Comparison of scoring methods