Open Access Research article

Determination of dosage compensation of the mammalian X chromosome by RNA-seq is dependent on analytical approach

Nathaniel K Jue, Michael B Murphy, Seth D Kasowitz, Sohaib M Qureshi, Craig J Obergfell, Sahar Elsisi, Robert J Foley, Rachel J O’Neill and Michael J O’Neill*

Author Affiliations

Department of Molecular and Cell Biology, University of Connecticut, 354 Mansfield Rd. U-2131, Storrs, CT 06235, USA

For all author emails, please log on.

BMC Genomics 2013, 14:150  doi:10.1186/1471-2164-14-150

Published: 6 March 2013

Additional files

Additional file 1: Table S1:

Proportion of total genes removed through implementation of methods from cited dosage compensation studies. Proportion of total genes that yielded a FPKM value > 0 that would not be included in the final calculations of RXE as defined by of the authors in the methods of the studies referenced below. a Utilizing described Miller’s Jackknife/Mann–Whitney U-test approach, applied unique mapping, established all genes that had a FPKM of zero, then removed a compensatory amount of genes from the upper end of the distribution. b Applied unique mapping, then removed all genes that had a FPKM <1. c Applied non-unique, splicing mapping, then removed all genes that had a FPKM <1. d Applied non-unique, splicing mapping, then removed outliers.

Format: ZIP Size: 261KB Download file

Open Data

Additional file 2: Table S2:

Paralog enrichment by chromosomal location in both mouse and human. Paralogous transcripts were determined by using BioMart (Ensembl), isolating paralogs with identity >70%. Number of genes per chromosome was calculated using RefSeq genome annotation.

Format: ZIP Size: 229KB Download file

Open Data

Additional file 3: Table S3:

RXE comparison using analysis methods implemented in cited dosage compensation studies. a Applied unique mapping. included all expressed genes and reported median values. b Applied unique mapping, then removed all genes that had a FPKM <1 and reported median values. c Applied non-unique, splicing mapping, then removed all genes that had a FPKM <1. Gene expression was calculated using cufflinks algorithm for all three analysis strategies to produce RXE values.

Format: ZIP Size: 177KB Download file

Open Data

Additional file 4: Table S4:

Reads and genes mapped by mapping parameter, for libraries analyzed in both mouse and human.

Format: ZIP Size: 328KB Download file

Open Data

Additional file 5: Figure S1:

Box plots of log2-transformed data of all FPKM values by chromosomal location. (A) human brain, (B) human liver, (C) normal human lymphoblast, (D) Xm human lymphoblast, (E) Xm mouse brain, (F) Xp mouse brain, (G) 40, XX mouse brain, (H) 40, XY mouse brain, (I) mouse brain.

Format: ZIP Size: 625KB Download file

Open Data

Additional file 6: Table S5:

Small RNA and riboprotein enrichment based on library preparation (Illumina or SOLiD). We compared small RNA (sno and micro) and riboprotein biases using three different mapping parameters: unique, non-unique, and non-unique splicing (NUS). Six different libraries were analyzed including: Xm lymphoblast (SOLiD), human brain (Illumina), human liver (Illumina) and Xm mouse brain (SOLiD), Xp mouse brain (SOLiD), and 40, XX mouse brain (SOLiD).

Format: ZIP Size: 188KB Download file

Open Data

Additional file 7: Figure S2:

Library size affects relative X-chromosome expression values in mammalian tissues. Plot of average log2-transformed RXE based on number of reads mapped. Includes data from human lymphoblast 45, Xm (n=1), human lymphoblast 45, Xp (n=1), human lymphoblast (n=10), human brain (n=1), and human liver (n=1) RNA-seq samples.

Format: ZIP Size: 255KB Download file

Open Data

Additional file 8: Table S6:

Gene counts for relative X-chromosome expression (RXE) values for cis and trans paralogs associated with GO terms binding activity and enzyme activity for five human tissue samples as described in Figure 4.

Format: ZIP Size: 241KB Download file

Open Data