Deep Learning (DL) has recently enabled unprecedented advances in one of the grand challenges in computational biology: the half-century-old problem of protein structure prediction. In this paper we discuss recent advances, limitations, and future perspectives of DL on five broad areas: protein structure prediction, protein function prediction, genome engineering, systems biology and data integration, and phylogenetic inference. We discuss each application area and cover the main bottlenecks of DL approaches, such as training data, problem scope, and the ability to leverage existing DL architectures in new contexts. To conclude, we provide a summary of the subject-specific and general challenges for DL across the biosciences.

We describe MetAMOS, an open source and modular metagenomic assembly and analysis pipeline. MetAMOS represents an important step towards fully automated metagenomic analysis, starting with next-generation sequencing reads and producing genomic scaffolds, open-reading frames and taxonomic or functional annotations. MetAMOS can aid in reducing assembly errors, commonly encountered when assembling metagenomic samples, and improves taxonomic assignment accuracy while also reducing computational cost. MetAMOS can be downloaded from: https://github.com/treangen/MetAMOS.

The rhesus macaque (Macaca mulatta) is a key species for advancing biomedical research. Like all draft mammalian genomes, the draft rhesus assembly (rheMac2) has gaps, sequencing errors and misassemblies that have prevented automated annotation pipelines from functioning correctly. Another rhesus macaque assembly, CR_1.0, is also available but is substantially more fragmented than rheMac2 with smaller contigs and scaffolds. Annotations for these two assemblies are limited in completeness and accuracy. High quality assembly and annotation files are required for a wide range of studies including expression, genetic and evolutionary analyses. We report a new de novo assembly of the rhesus macaque genome (MacaM) that incorporates both the original Sanger sequences used to assemble rheMac2 and new Illumina sequences from the same animal. MacaM has a weighted average (N50) contig size of 64 kilobases, more than twice the size of the rheMac2 assembly and almost five times the size of the CR_1.0 assembly. The MacaM chromosome assembly incorporates information from previously unutilized mapping data and preliminary annotation of scaffolds. Independent assessment of the assemblies using Ion Torrent read alignments indicates that MacaM is more complete and accurate than rheMac2 and CR_1.0. We assembled messenger RNA sequences from several rhesus tissues into transcripts which allowed us to identify a total of 11,712 complete proteins representing 9,524 distinct genes. Using a combination of our assembled rhesus macaque transcripts and human transcripts, we annotated 18,757 transcripts and 16,050 genes with complete coding sequences in the MacaM assembly. Further, we demonstrate that the new annotations provide greatly improved accuracy as compared to the current annotations of rheMac2. Finally, we show that the MacaM genome provides an accurate resource for alignment of reads produced by RNA sequence expression studies. The MacaM assembly and annotation files provide a substantially more complete and accurate representation of the rhesus macaque genome than rheMac2 or CR_1.0 and will serve as an important resource for investigators conducting next-generation sequencing studies with nonhuman primates. This article was reviewed by Dr. Lutz Walter, Dr. Soojin Yi and Dr. Kateryna Makova.

In order to determine the role of the database in taxonomic sequence classification, we examine the influence of the database over time on k-mer-based lowest common ancestor taxonomic classification. We present three major findings: the number of new species added to the NCBI RefSeq database greatly outpaces the number of new genera; as a result, more reads are classified with newer database versions, but fewer are classified at the species level; and Bayesian-based re-estimation mitigates this effect but struggles with novel genomes. These results suggest a need for new classification approaches specially adapted for large databases.

Motivation: Since 2016, the number of microbial species with available reference genomes in NCBI has more than tripled. Multiple genome alignment, the process of identifying nucleotides across multiple genomes which share a common ancestor, is used as the input to numerous downstream comparative analysis methods. Parsnp is one of the few multiple genome alignment methods able to scale to the current era of genomic data; however, there has been no major release since its initial release in 2014. Results: To address this gap, we developed Parsnp v2, which significantly improves on its original release. Parsnp v2 provides users with more control over executions of the program, allowing Parsnp to be better tailored for different use-cases. We introduce a partitioning option to Parsnp, which allows the input to be broken up into multiple parallel alignment processes which are then combined into a final alignment. The partitioning option can reduce memory usage by over 4× and reduce runtime by over 2×, all while maintaining a precise core-genome alignment. The partitioning workflow is also less susceptible to complications caused by assembly artifacts and minor variation, as alignment anchors only need to be conserved within their partition and not across the entire input set. We highlight the performance on datasets involving thousands of bacterial and viral genomes. Availability and implementation: Parsnp v2 is available at https://github.com/marbl/parsnp.