Ranked Choice Voting for Representative Transcripts with TRaCE

Genome sequencing projects annotate protein-coding gene models with multiple transcripts, aiming to represent all of the available transcript evidence. However, downstream analyses often operate on only one representative transcript per gene locus, sometimes known as the canonical transcript. To choose canonical transcripts, TRaCE (Transcript Ranking and Canonical Election) holds an 'election' in which a set of RNA-seq samples rank transcripts by annotation edit distance. These sample-specific votes are tallied along with other criteria such as protein length and InterPro domain coverage. The winner is selected as the canonical transcript, but the election proceeds through multiple rounds of voting to order all the transcripts by relevance. Based on the set of expression data provided, TRaCE can identify the most common isoforms from a broad expression atlas or prioritize alternative transcripts expressed in specific contexts.

Automatic annotation of Cryo-EM maps with the convolutional neural network Haruspex

In recent years, three-dimensional density maps reconstructed from single particle images obtained by electron cryo-microscopy (Cryo-EM) have reached unprecedented resolution. However, map interpretation can be challenging, in particular if the constituting structures require de-novo model building or are very mobile. Here, we demonstrate the potential of convolutional neural networks for the annotation of Cryo-EM maps: our network Haruspex has been trained on a carefully curated set of 293 experimentally derived reconstruction maps to automatically annotate protein secondary structure elements as well as RNA/DNA. It can be straightforwardly applied to annotate newly reconstructed maps to support domain placement or to supply a starting point for main-chain placement. Due to its high recall and precision rates of 95.1% and 80.3%, respectively, on an independent test set of 122 maps, it can also be used for validation during model building. The trained network will be available as part of the CCP-EM suite.