On 2015 Feb 10, Pieter-Jan Volders commented:

At the time of submission, we were unaware of a specific RefSeq lncRNA subset. As a result, we initially used the NR_* records larger than 200 nucleotides. When dr. Kim D. Pruitt contacted us regarding this issue, we repeated the analysis for the suggested RefSeq subset. This subset contains 4774 transcripts and was obtained through the UCSC table browser. As expected, the percentage of transcripts passing the PhyloCSF cutoff has decreased from 48% to 14% of the RefSeq subset. The online manuscript was updated and the current Figure 3 (online as of January 15, 2015) represents these new results. Additionally, LNCipedia.org was updated as well, and the RefSeq records that do not represent lncRNAs were removed from the database version 3.1.

It is worth noting that we could not find any information on the keyword that was used in the query suggested by dr. Kimm D. Pruit (biomolncrnalncrna), neither on the RefSeq website, nor in the cited manuscript. It is unclear to us how many researchers are aware of this and we would like to suggest to RefSeq to indicate this subset on their website as it is of great value to the lncRNA research community.

On 2014 Dec 03, Kim D Pruitt commented:

The RefSeq results that are presented in Figure 3 do not accurately reflect the high quality of the RefSeq lncRNA dataset.

The article does not describe how this dataset was defined but the authors kindly provided this information as well as a file listing all of the RefSeq accessions and PhyloCSF results. The authors’ definition of RefSeq lncRNA was too simple as it was primarily based on the RefSeq accession prefix (‘NR_’). However, this accession prefix is used by the RefSeq project for several types of noncoding transcripts (PMID:18927115). Roughly 50% of the RefSeq dataset analyzed represent transcribed pseudogenes or noncoding transcripts for protein-coding genes.

On 2014 Dec 03, Kim D Pruitt commented:

The RefSeq results that are presented in Figure 3 do not accurately reflect the high quality of the RefSeq lncRNA dataset.

The article does not describe how this dataset was defined but the authors kindly provided this information as well as a file listing all of the RefSeq accessions and PhyloCSF results. The authors’ definition of RefSeq lncRNA was too simple as it was primarily based on the RefSeq accession prefix (‘NR_’). However, this accession prefix is used by the RefSeq project for several types of noncoding transcripts (PMID:18927115). Roughly 50% of the RefSeq dataset analyzed represent transcribed pseudogenes or noncoding transcripts for protein-coding genes.