Skip to main content
BMC Bioinformatics logoLink to BMC Bioinformatics
. 2019 Dec 20;20:727. doi: 10.1186/s12859-019-3305-4

Correction to: Detailed prediction of protein sub-nuclear localization

Maria Littmann 1,✉,#, Tatyana Goldberg 1,#, Sebastian Seitz 1, Mikael Bodén 2, Burkhard Rost 1,3,4,5
PMCID: PMC6925513  PMID: 31861997

Correction to: BMC Bioinformatics (2019) 20:205

https://doi.org/10.1186/s12859-019-2790-9

Following publication of the original article [1], the author reported that an incorrect figure has been published as Fig. 2. The correct Fig. 2 is shown below.

Fig. 2.

Fig. 2

Highly reliable predictions more accurate. The reliability index (RI, x-axis) of LocNuclei scaled between 0 (unreliable) and 100 (reliable). It related the prediction strength to the performance. The data for this figure were binned in intervals of 20. Each point reflected the cumulative performance, i.e. we computed accuracy (Q13 and Q2) and coverage (percentage of proteins for which predictions were made above given RI). a For the prediction of 13 nuclear sub-structures, 19% of all proteins were predicted at RI > 60 (point marked by dotted lines). For this top 19%, accuracy rose from the average Q13 = 62% (indicated by leftmost black point) to 75% (point marked by dotted lines). For our data set, RI < 20 did not correlate with accuracy. b For the prediction of traveler proteins, 29% of all proteins were predicted at RI > 60 (part B, point marked by dotted lines) with Q2 = 78% (point marked by dotted lines, improving over the average of 72% by six percentage points)

The publisher apologizes to the authors and readers for the inconvenience.

Footnotes

Maria Littmann and Tatyana Goldberg contributed equally to this work.

Reference


Articles from BMC Bioinformatics are provided here courtesy of BMC

RESOURCES