​Supplemental Material for Basu and Larsson, 2018

Supplementary figure
1. (S1_G3.pdf) General statistics and pipeline. A) Number of primary tumor
samples for which raw sequencing data was downloaded, processed and analyzed.
B) Pipeline for mapping and quantification of sequencing reads on the human
genome.  

Supplementary figure
2. (S2_G3.pdf) Distribution of spearman correlation coefficients comparing the
expression of lncRNAs and proximal coding genes (100 Kb windows) across tumor
samples in 32 cancer types. 

Supplementary figure
3. (S3_G3.pdf) Distribution of spearman correlation coefficients comparing the
expression of lncRNAs and proximal coding genes (100 Kb windows) across tumor
samples in 32 cancer types where coding and lncRNA genomic coordinates are shuffled. 

Supplementary figure
4. (S4_G3.pdf) Expression correlation enrichment and distribution at a
stringent P cut-off A) Enrichment of significant correlations (P < 1e-5) in
between lncRNAs and neighbor coding genes (100 Kb) compared to random lncRNA-coding
pairs. B) Distribution of spearman correlation coefficients for coding genes
and their proximal lncRNAs and proximal coding genes (100 Kb, ρ, P < 1e-5).

Supplementary figure
5. (S5_G3.pdf) Percentage of coding/lncRNA pairs showing positive and negative
correlation at different distance threshold across multiple cancer types. A) 25
Kb B) 50 Kb C) 100 Kb D) mean percentage of positive and negative correlations
across all cancers at 25, 50 and 100 Kb. 

Supplementary figure
6. (S6_G3.pdf) Cytoplasmic/Nucleus Relative Concentration Index distribution in
coding/lncRNA pairs showing negative (LCN) and positive (LCP) correlation
compared against all lncRNAs. Grey asterisk marks a significant difference in
distribution between LCN and all lncRNA while red asterisk marks the same
between LCN and LCP (Student’s t-test, P < 0.05).

Supplementary table 1.
(Supplementary_table-1.xls) Known long non-coding RNA and coding gene
associations captured by expression correlation analysis across multiple cancer
types.

Supplementary table 2.
(Supplementary_table-2.xls) 193 pairs of lncRNAs and coding genes (involving
178 unique lncRNAs and 188 unique coding genes) showing negative expression
correlation in the majority of cancers with significant correlation (P < 1e-5)
in at least one cancer type.

Supplementary table 3.
(Supplementary_table-3.xls) Predicted subcellular localization of lncRNAs which
show negative expression correlation in the majority of cancers with
significant correlation (P < 1e-5) in at least one cancer type.

Supplementary table 4.
(Supplementary_table-4.xls) LncRNAs which are significantly negatively
correlated with a coding gene at the expression level while the lncRNA
expression correlates positively with the methylation levels at the
transcription start site (TSS) of their coding counterparts.