Limits...
A method for detecting long non-coding RNAs with tiled RNA expression microarrays.

Lund SH, Gudbjartsson DF, Rafnar T, Sigurdsson A, Gudjonsson SA, Gudmundsson J, Stefansson K, Stefansson G - PLoS ONE (2014)

Bottom Line: The robustness of the method is tested by utilizing repeated copies of tiled probes.The method shows high consistency between experiments that used the same samples, but different probe layout.There also is statistically significant consistency when comparing experiments with different samples.

View Article: PubMed Central - PubMed

Affiliation: Faculty of Physical Sciences, University of Iceland, Reykjavik, Iceland.

ABSTRACT
Long non-coding ribonucleic acids (lncRNAs) have been proposed as biomarkers in prostate cancer. This paper proposes a selection method which uses data from tiled microarrays to identify relatively long regions of moderate expression independent of the microarray platform and probe design. The method is used to search for candidate long non-coding ribonucleic acids (lncRNAs) at locus 8q24 and is run on three independent experiments which all use samples from prostate cancer patients. The robustness of the method is tested by utilizing repeated copies of tiled probes. The method shows high consistency between experiments that used the same samples, but different probe layout. There also is statistically significant consistency when comparing experiments with different samples. The method selected the long non-coding ribonucleic acid PCNCR1 in all three experiments.

Show MeSH

Related in: MedlinePlus

The number of Monte Carlo simulations for which each region is chosen by the selection method using arrays with different samples.The genomic location of the regions on 8q24 is on the x-axis. The proportion of Monte Carlo simulations for which the region was chosen is on the y-axis. The graph is shown with two different colourings, representing whether the region was among the previously experiment-wise selected regions (cyan) or not (pink). Those who were selected previously in Experiment 1 are shown at the top graph, Experiment 2 in the middle and Experiment 3 at the bottom. The simulations are done on the ten repeated spots for each probe for all nine arrays in Experiment 3.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4061049&req=5

pone-0099899-g003: The number of Monte Carlo simulations for which each region is chosen by the selection method using arrays with different samples.The genomic location of the regions on 8q24 is on the x-axis. The proportion of Monte Carlo simulations for which the region was chosen is on the y-axis. The graph is shown with two different colourings, representing whether the region was among the previously experiment-wise selected regions (cyan) or not (pink). Those who were selected previously in Experiment 1 are shown at the top graph, Experiment 2 in the middle and Experiment 3 at the bottom. The simulations are done on the ten repeated spots for each probe for all nine arrays in Experiment 3.

Mentions: Figure 3 shows a graph of the proportion of Monte Carlo simulations for which each region was chosen among the top 25. The colouring indicates whether the region was among the 25 experiment-wise selected regions for Experiments 1, 2 and 3. It is seen that the majority of regions are never among the top 25, whereas 14 regions are selected in at least 75% of the simulations. The experiment-wise selected regions seem to be selected more often in the Monte-Carlo simulation.


A method for detecting long non-coding RNAs with tiled RNA expression microarrays.

Lund SH, Gudbjartsson DF, Rafnar T, Sigurdsson A, Gudjonsson SA, Gudmundsson J, Stefansson K, Stefansson G - PLoS ONE (2014)

The number of Monte Carlo simulations for which each region is chosen by the selection method using arrays with different samples.The genomic location of the regions on 8q24 is on the x-axis. The proportion of Monte Carlo simulations for which the region was chosen is on the y-axis. The graph is shown with two different colourings, representing whether the region was among the previously experiment-wise selected regions (cyan) or not (pink). Those who were selected previously in Experiment 1 are shown at the top graph, Experiment 2 in the middle and Experiment 3 at the bottom. The simulations are done on the ten repeated spots for each probe for all nine arrays in Experiment 3.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4061049&req=5

pone-0099899-g003: The number of Monte Carlo simulations for which each region is chosen by the selection method using arrays with different samples.The genomic location of the regions on 8q24 is on the x-axis. The proportion of Monte Carlo simulations for which the region was chosen is on the y-axis. The graph is shown with two different colourings, representing whether the region was among the previously experiment-wise selected regions (cyan) or not (pink). Those who were selected previously in Experiment 1 are shown at the top graph, Experiment 2 in the middle and Experiment 3 at the bottom. The simulations are done on the ten repeated spots for each probe for all nine arrays in Experiment 3.
Mentions: Figure 3 shows a graph of the proportion of Monte Carlo simulations for which each region was chosen among the top 25. The colouring indicates whether the region was among the 25 experiment-wise selected regions for Experiments 1, 2 and 3. It is seen that the majority of regions are never among the top 25, whereas 14 regions are selected in at least 75% of the simulations. The experiment-wise selected regions seem to be selected more often in the Monte-Carlo simulation.

Bottom Line: The robustness of the method is tested by utilizing repeated copies of tiled probes.The method shows high consistency between experiments that used the same samples, but different probe layout.There also is statistically significant consistency when comparing experiments with different samples.

View Article: PubMed Central - PubMed

Affiliation: Faculty of Physical Sciences, University of Iceland, Reykjavik, Iceland.

ABSTRACT
Long non-coding ribonucleic acids (lncRNAs) have been proposed as biomarkers in prostate cancer. This paper proposes a selection method which uses data from tiled microarrays to identify relatively long regions of moderate expression independent of the microarray platform and probe design. The method is used to search for candidate long non-coding ribonucleic acids (lncRNAs) at locus 8q24 and is run on three independent experiments which all use samples from prostate cancer patients. The robustness of the method is tested by utilizing repeated copies of tiled probes. The method shows high consistency between experiments that used the same samples, but different probe layout. There also is statistically significant consistency when comparing experiments with different samples. The method selected the long non-coding ribonucleic acid PCNCR1 in all three experiments.

Show MeSH
Related in: MedlinePlus