Email updates

Keep up to date with the latest news and content from Genome Biology and BioMed Central.

Open Access Research

Analysis of expressed SNPs identifies variable extents of expression from the human inactive X chromosome

Allison M Cotton1, Bing Ge2, Nicholas Light2, Veronique Adoue2, Tomi Pastinen2* and Carolyn J Brown1*

Author Affiliations

1 Department of Medical Genetics, Molecular Epigenetics Group, Life Sciences Institute, University of British Columbia, Vancouver, BC V6T 1Z3, Canada

2 McGill University and Genome Québec Innovation Centre, Department of Human and Medical Genetics, McGill University, Montréal, Québec H3A 0G1, Canada

For all author emails, please log on.

Genome Biology 2013, 14:R122  doi:10.1186/gb-2013-14-11-r122

Published: 1 November 2013

Additional files

Additional file 1: Table S1:

Subject training set genes. List of 177 genes in the subject training set. All genes were previously found to be subject to XCI by expression analysis in somatic cell hybrids [10] (expression in 0 to 22% of examined somatic cell hybrids) and were also subject to XCI based on DNA methylation analysis in all tissues [13]. The expression of a subset of genes, marked with an asterisk, as examined in fibroblasts [10] also supports that these genes are subject to XCI (average of less than 10% Xi expression).

Format: XLSX Size: 10KB Download file

Open Data

Additional file 2: Figure S1:

Identification of females with skewed XCI. Conversion of AI into XCI status in a group 2 female with a high degree of skewing (A) compared to a female with low skewing (B). The line demonstrates the linear regression between the female analyzed and the average AI from the CEU group 1 females (subject and escape genes only). The horizontal shading denotes the ranges of AI that correspond to the XCI statuses in the group 2 female: dark green (E1), bright green (E2), light green (E3) or red (subject to XCI). The lower degree of skewing of XCI (B) results in a condensed range of escape from XCI. For all group 2 females, the boundary between E3 and S was determined using the AI at which there was 10% expression from the Xi once corrected for skewing. A complete list of boundaries can be found in Additional file 5. (C) The linear regression between the average group 1 female AI and group R females was not significant and therefore AIs in group R females could not be converted in XCI statuses. A complete list of group R females can be found in Additional file 5.

Format: PNG Size: 53KB Download file

Open Data

Additional file 3: Table S2:

Escape training set genes. List of 43 genes in the escape training set. All genes were previously found to be escape from XCI by expression analysis in somatic cell hybrids [10] (expression in 78 to 100% of examined somatic cell hybrids) and also escaped from XCI based on DNA methylation analysis in all tissues [13]. The expression of a subset of genes, marked with an asterisk, as examined in fibroblasts [10] also supports that these genes escape from XCI (average of more than 10% Xi expression). Those genes located in the PAR1 are marked with a pound sign.

Format: XLSX Size: 10KB Download file

Open Data

Additional file 4: Figure S2:

Minimum cDNA probe intensity thresholds differ in each sample set. (A,B) CEU sample set, (C,D) YRI sample set, (E,F) FIB sample set. (A,C,E) The average expression (both cDNA channels) and AI were determined for all uninformative females (CEU, n = 30; YRI, n = 31; FIB, n = 38), then graphed and a one phase decay linear regression performed. The Tau for each population was then determined (solid black line in (B,D,F)) and the 95% confidence interval also plotted (dotted black line in (B,D,F)). Only probes with a total cDNA expression greater than Tau were used in further analysis. Details of Tau thresholds are in Additional files 6 and 11.

Format: PNG Size: 116KB Download file

Open Data

Additional file 5: Table S3:

Thresholds of AI for conversion into XCI status following linear regression in non-group 1 females. Group 2 females show a significant linear regression with the average AI from the group 1 females whereas group R females are not significant and therefore AIs cannot be converted into XCI status using the slope of the linear regression line.

Format: XLSX Size: 14KB Download file

Open Data

Additional file 6: Table S4:

The majority of probes are removed due to a low cDNA probe intensity.

Format: XLSX Size: 9KB Download file

Open Data

Additional file 9: Table S6:

AI thresholds used to translate XCI status for group 1 females from each sample set. Within each sample set the E1:E2 and E2:E3 boundaries are the same for all individuals while the E3:S boundary differs based on the degree of skewing in each female. The AI that corresponded to 10% Xi expression was calculated based on the degree of skewing in each female. The AI for each group 1 female can be found in Additional file 12.

Format: XLSX Size: 10KB Download file

Open Data

Additional file 12: Table S8:

AI at 10% Xi expression for each group 1 female.

Format: XLSX Size: 10KB Download file

Open Data

Additional file 7: Table S5:

List of all genic XCI statuses determined. The gene name, if the gene is a member of the escape or subject training set, the total number of informative females, the percentage of which escape from and the genic XCI status are listed along with the average %Xi and the standard deviation of the %Xi for each gene. Genic XCI status cells are colored, subject to XCI (red), variable escape from XCI (purple) and escape from XCI (green). Genes are listed from Xp to Xq.

Format: XLSX Size: 60KB Download file

Open Data

Additional file 8: Figure S3:

Histone ChIP AI of combined histone modifications is highest at genes subject to XCI. Error bars represent the standard error of the mean while the color indicates the assigned genic XCI status. Genes that escape XCI and have a %Xi expression within the PAR1 range (dark green), genes that escape XCI and have a %Xi expression outside the PAR1 range (green), genes that are subject to XCI and have a %Xi expression greater than 5% Xi (red) and genes that are subject to XCI and have a %Xi expression less than 5% Xi (dark red). Significant differences between means are shown as asterisks (*P-value 0.05 to 1.0 e-5, **P-value 1.0 e-5 to 1.0 e-15, ***P-value <1.0 e-15). All P-values were corrected for multiple comparisons.

Format: PNG Size: 18KB Download file

Open Data

Additional file 13: Figure S5:

Decision tree to determine XCI status across sample sets. In order to compare XCI status between the three sample sets a standard set of yes/no questions was devised. To begin (black rectangle in the center) the two LCL sample sets (CEU and YRI) are examined, then the FIB sample set is brought in to determine if differences in XCI status were the result of population or cell line differences. In total, six different cross-sample set XCI statuses were defined: subject in all sample sets (red), VE (variable escape) in all sample sets (purple), escape in all sample sets (green), population-specific XCI (orange), cell line-specific XCI (blue), population and cell line-specific XCI (blue and orange stripes) and inconsistent XCI between samples sets (white).

Format: PNG Size: 63KB Download file

Open Data

Additional file 10: Figure S4:

Allelic expression analysis in group R females reveals no evidence for X-linked imprinting. (A) Allelic expression bias in individual informative females with each column representing one gene (only genes with at least one informative female are included) and each row a single female (group R females only) from the three sample sets. The allelic expression bias of each female is either mono-allelic (red) or bi-allelic (green). The genic expression status was determined by calculating the percentage of females that were bi-allelic for each gene: 0 to 22% of informative females bi-allelic, genic bias is mono-allelic (red); 22 to 78% of informative females bi-allelic, genic bias is variable allelic (purple); or 78 to 100% of informative females bi-allelic, genic bias is bi-allelic (green). (B) Distribution of AIs observed for every gene with at least one female with mono-allelic expression. For each gene, informative females are represented with a different shape based on the sample set (CEU, square; YRI, circle; FIB, triangle) and color based on allelic expression status (red, mono-allelic; green, bi-allelic). Below the genic allelic-expression status is given (bi = bi-allelic, VA = variable allelic, mono = mono-allelic). MAP7D2, the previously reported X-linked gene, is shown to the far right.

Format: PNG Size: 236KB Download file

Open Data

Additional file 11: Table S7:

Minimum probe intensities are comparable between males and females. Tau was also determined for all uninformative probes in males for each sample set.

Format: XLSX Size: 9KB Download file

Open Data