Orthology Prediction Research Articles

PorthoMCL: Parallel orthology prediction using MCL for the realm of massive genome availability

BackgroundFinding orthologous genes among multiple sequenced genomes is a primary step in comparative genomics studies. With the number of sequenced genomes increasing exponentially, comparative genomics becomes more powerful than ever for genomic analysis. However, the very large number of genomes in need of analysis makes conventional orthology prediction methods incapable of this task. Thus, an ultrafast tool is urgently needed.ResultsHere, we present PorthoMCL, a fast tool for finding orthologous genes among a very large number of genomes. PorthoMCL can be run on a single machine or in parallel on computer clusters. We have demonstrated PorthoMCL’s capability by identifying orthologs in 2,758 prokaryotic genomes. The results are available for download at: http://ehsun.me/go/porthomcl/.ConclusionsPorthoMCL is a fast and easy to run tool for identifying orthology among any number of genomes with minimal requirements. PorthoMCL will facilitate comparative genomics analysis with increasing number of available genomes thanks to the rapidly evolving sequencing technologies.

Transcriptologs: A Transcriptome-Based Approach to Predict Orthology Relationships.

The detection of orthologs is a key approach in genomics, useful to understand gene evolution and phylogenetic relationships and essential for gene function prediction. However, a reliable annotation of the encoded protein regions is still a limiting aspect in genomics, mainly due to the lack of confirmatory experimental evidence at proteome level. Nevertheless, the current ortholog collections are generally based on protein sequence comparisons, in addition to the availability of large transcriptome sequence collections. We developed Transcriptologs, a method for the prediction of orthologs based on similarities of translated fragments from messenger RNAs of 2 species. We implemented a procedure to extend BLAST-based alignments and to define orthologs based on the Bidirectional Best Hit approach. Results from a test case on Arabidopsis thaliana and Sorghum bicolor transcript collections revealed in some cases outperformance of Transcriptologs in comparison with a classical protein-based analysis in terms of alignment quality, revealing similarities otherwise not detectable.

WORMHOLE: Novel Least Diverged Ortholog Prediction through Machine Learning.

The rapid advancement of technology in genomics and targeted genetic manipulation has made comparative biology an increasingly prominent strategy to model human disease processes. Predicting orthology relationships between species is a vital component of comparative biology. Dozens of strategies for predicting orthologs have been developed using combinations of gene and protein sequence, phylogenetic history, and functional interaction with progressively increasing accuracy. A relatively new class of orthology prediction strategies combines aspects of multiple methods into meta-tools, resulting in improved prediction performance. Here we present WORMHOLE, a novel ortholog prediction meta-tool that applies machine learning to integrate 17 distinct ortholog prediction algorithms to identify novel least diverged orthologs (LDOs) between 6 eukaryotic species—humans, mice, zebrafish, fruit flies, nematodes, and budding yeast. Machine learning allows WORMHOLE to intelligently incorporate predictions from a wide-spectrum of strategies in order to form aggregate predictions of LDOs with high confidence. In this study we demonstrate the performance of WORMHOLE across each combination of query and target species. We show that WORMHOLE is particularly adept at improving LDO prediction performance between distantly related species, expanding the pool of LDOs while maintaining low evolutionary distance and a high level of functional relatedness between genes in LDO pairs. We present extensive validation, including cross-validated prediction of PANTHER LDOs and evaluation of evolutionary divergence and functional similarity, and discuss future applications of machine learning in ortholog prediction. A WORMHOLE web tool has been developed and is available at http://wormhole.jax.org/.

Expression Pattern Similarities Support the Prediction of Orthologs Retaining Common Functions after Gene Duplication Events.

The identification of functionally equivalent, orthologous genes (functional orthologs) across genomes is necessary for accurate transfer of experimental knowledge from well-characterized organisms to others. This frequently relies on automated, coding sequence-based approaches such as OrthoMCL, Inparanoid, and KOG, which usually work well for one-to-one homologous states. However, this strategy does not reliably work for plants due to the occurrence of extensive gene/genome duplication. Frequently, for one query gene, multiple orthologous genes are predicted in the other genome, and it is not clear a priori from sequence comparison and similarity which one preserves the ancestral function. We have studied 11 organ-dependent and stress-induced gene expression patterns of 286 Arabidopsis lyrata duplicated gene groups and compared them with the respective Arabidopsis (Arabidopsis thaliana) genes to predict putative expressologs and nonexpressologs based on gene expression similarity. Promoter sequence divergence as an additional tool to substantiate functional orthology only partially overlapped with expressolog classification. By cloning eight A. lyrata homologs and complementing them in the respective four Arabidopsis loss-of-function mutants, we experimentally proved that predicted expressologs are indeed functional orthologs, while nonexpressologs or nonfunctionalized orthologs are not. Our study demonstrates that even a small set of gene expression data in addition to sequence homologies are instrumental in the assignment of functional orthologs in the presence of multiple orthologs.

Simultaneous prediction of enzyme orthologs from chemical transformation patterns for de novo metabolic pathway reconstruction.

Motivation: Metabolic pathways are an important class of molecular networks consisting of compounds, enzymes and their interactions. The understanding of global metabolic pathways is extremely important for various applications in ecology and pharmacology. However, large parts of metabolic pathways remain unknown, and most organism-specific pathways contain many missing enzymes. Results: In this study we propose a novel method to predict the enzyme orthologs that catalyze the putative reactions to facilitate the de novo reconstruction of metabolic pathways from metabolome-scale compound sets. The algorithm detects the chemical transformation patterns of substrate–product pairs using chemical graph alignments, and constructs a set of enzyme-specific classifiers to simultaneously predict all the enzyme orthologs that could catalyze the putative reactions of the substrate–product pairs in the joint learning framework. The originality of the method lies in its ability to make predictions for thousands of enzyme orthologs simultaneously, as well as its extraction of enzyme-specific chemical transformation patterns of substrate–product pairs. We demonstrate the usefulness of the proposed method by applying it to some ten thousands of metabolic compounds, and analyze the extracted chemical transformation patterns that provide insights into the characteristics and specificities of enzymes. The proposed method will open the door to both primary (central) and secondary metabolism in genomics research, increasing research productivity to tackle a wide variety of environmental and public health matters.Availability and Implementation:Contact: maskot@bio.titech.ac.jp

UpCoT: an integrated pipeline tool for clustering upstream DNA sequences of orthologous genes in prokaryotic genomes.

UpCoT is a pipeline tool developed by automating the series of steps involved in prediction of cis-regulatory elements. UpCoT generates orthologs for each gene in target genome using bi-directional best blast hit against the reference genomes, then identifies potential orthologous transcriptional units using intergenic distance. Finally it generates the FASTA files containing upstream sequences of orthologous transcriptional units of each gene in target genome. The inputs of UpCoT are protein sequence files (*.faa), genome sequence files (*.fna) and gene co-ordinate files (*.ptt) for target and reference genomes. The clustered-upstream DNA sequences can be used by motif prediction tool, such as MEME, Bio-prospector, Gibbs motif sampler, MDscan for prediction of conserved DNA elements. We tested the performance of UpCoT by selecting the genome of Synechocystis sp PCC 6803 as the target and 13 different cyanobacterial genomes as reference. The clustered upstream sequences generated by UpCoT of groES, ycf24 and nirA were used for cis-regulatory element prediction. The results were consistent with the experimentally identified cis-regulatory elements. Therefore, UpCoT is a reliable and automated pipeline package for prediction of orthologs, orthologous transcriptional units, and orthologous upstream sequences of a selected prokaryotic genome. UpCoT can be downloaded from http://jssplab.uohyd.ac.in/upcot/.

Survey of Genes Involved in Biosynthesis, Transport, and Signaling of Phytohormones with Focus onSolanum lycopersicum

Phytohormones control the development and growth of plants, as well as their response to biotic and abiotic stress. The seven most well-studied phytohormone classes defined today are as follows: auxins, ethylene, cytokinin, abscisic acid, jasmonic acid, gibberellins, and brassinosteroids. The basic principle of hormone regulation is conserved in all plants, but recent results suggest adaptations of synthesis, transport, or signaling pathways to the architecture and growth environment of different plant species. Thus, we aimed to define the extent to which information from the model plant Arabidopsis thaliana is transferable to other plants such as Solanum lycopersicum. We extracted the co-orthologues of genes coding for major pathway enzymes in A. thaliana from the translated genomes of 12 species from the clade Viridiplantae. Based on predicted domain architecture and localization of the identified proteins from all 13 species, we inspected the conservation of phytohormone pathways. The comparison was complemented by expression analysis of (co-) orthologous genes in S. lycopersicum. Altogether, this information allowed the assignment of putative functional equivalents between A. thaliana and S. lycopersicum but also pointed to some variations between the pathways in eudicots, monocots, mosses, and green algae. These results provide first insights into the conservation of the various phytohormone pathways between the model system A. thaliana and crop plants such as tomato. We conclude that orthologue prediction in combination with analysis of functional domain architecture and intracellular localization and expression studies are sufficient tools to transfer information from model plants to other plant species. Our results support the notion that hormone synthesis, transport, and response for most part of the pathways are conserved, and species-specific variations can be found.

Beyond the E-Value: Stratified Statistics for Protein Domain Prediction.

E-values have been the dominant statistic for protein sequence analysis for the past two decades: from identifying statistically significant local sequence alignments to evaluating matches to hidden Markov models describing protein domain families. Here we formally show that for “stratified” multiple hypothesis testing problems—that is, those in which statistical tests can be partitioned naturally—controlling the local False Discovery Rate (lFDR) per stratum, or partition, yields the most predictions across the data at any given threshold on the FDR or E-value over all strata combined. For the important problem of protein domain prediction, a key step in characterizing protein structure, function and evolution, we show that stratifying statistical tests by domain family yields excellent results. We develop the first FDR-estimating algorithms for domain prediction, and evaluate how well thresholds based on q-values, E-values and lFDRs perform in domain prediction using five complementary approaches for estimating empirical FDRs in this context. We show that stratified q-value thresholds substantially outperform E-values. Contradicting our theoretical results, q-values also outperform lFDRs; however, our tests reveal a small but coherent subset of domain families, biased towards models for specific repetitive patterns, for which weaknesses in random sequence models yield notably inaccurate statistical significance measures. Usage of lFDR thresholds outperform q-values for the remaining families, which have as-expected noise, suggesting that further improvements in domain predictions can be achieved with improved modeling of random sequences. Overall, our theoretical and empirical findings suggest that the use of stratified q-values and lFDRs could result in improvements in a host of structured multiple hypothesis testing problems arising in bioinformatics, including genome-wide association studies, orthology prediction, and motif scanning.

Genome Evolution and Innovation across the Four Major Lineages of Cryptococcus gattii.

ABSTRACTCryptococcus gattii is a fungal pathogen of humans, causing pulmonary infections in otherwise healthy hosts. To characterize genomic variation among the four major lineages of C. gattii (VGI, -II, -III, and -IV), we generated, annotated, and compared 16 de novo genome assemblies, including the first for the rarely isolated lineages VGIII and VGIV. By identifying syntenic regions across assemblies, we found 15 structural rearrangements, which were almost exclusive to the VGI-III-IV lineages. Using synteny to inform orthology prediction, we identified a core set of 87% of C. gattii genes present as single copies in all four lineages. Remarkably, 737 genes are variably inherited across lineages and are overrepresented for response to oxidative stress, mitochondrial import, and metal binding and transport. Specifically, VGI has an expanded set of iron-binding genes thought to be important to the virulence of Cryptococcus, while VGII has expansions in the stress-related heat shock proteins relative to the other lineages. We also characterized genes uniquely absent in each lineage, including a copper transporter absent from VGIV, which influences Cryptococcus survival during pulmonary infection and the onset of meningoencephalitis. Through inclusion of population-level data for an additional 37 isolates, we identified a new transcontinental clonal group that we name VGIIx, mitochondrial recombination between VGII and VGIII, and positive selection of multidrug transporters and the iron-sulfur protein aconitase along multiple branches of the phylogenetic tree. Our results suggest that gene expansion or contraction and positive selection have introduced substantial variation with links to mechanisms of pathogenicity across this species complex.

Quest for Orthologs Entails Quest for Tree of Life: In Search of the Gene Stream.

Quest for Orthologs (QfO) is a community effort with the goal to improve and benchmark orthology predictions. As quality assessment assumes prior knowledge on species phylogenies, we investigated the congruency between existing species trees by comparing the relationships of 147 QfO reference organisms from six Tree of Life (ToL)/species tree projects: The National Center for Biotechnology Information (NCBI) taxonomy, Opentree of Life, the sequenced species/species ToL, the 16S ribosomal RNA (rRNA) database, and trees published by Ciccarelli et al. (Ciccarelli FD, et al. 2006. Toward automatic reconstruction of a highly resolved tree of life. Science 311:1283–1287) and by Huerta-Cepas et al. (Huerta-Cepas J, Marcet-Houben M, Gabaldon T. 2014. A nested phylogenetic reconstruction approach provides scalable resolution in the eukaryotic Tree Of Life. PeerJ PrePrints 2:223) Our study reveals that each species tree suggests a different phylogeny: 87 of the 146 (60%) possible splits of a dichotomous and rooted tree are congruent, while all other splits are incongruent in at least one of the species trees. Topological differences are observed not only at deep speciation events, but also within younger clades, such as Hominidae, Rodentia, Laurasiatheria, or rosids. The evolutionary relationships of 27 archaea and bacteria are highly inconsistent. By assessing 458,108 gene trees from 65 genomes, we show that consistent species topologies are more often supported by gene phylogenies than contradicting ones. The largest concordant species tree includes 77 of the QfO reference organisms at the most. Results are summarized in the form of a consensus ToL (http://swisstree.vital-it.ch/species_tree) that can serve different benchmarking purposes.

Phylogenetic Method for High-Throughput Ortholog Detection

Accurate detection of orthologous proteins is a key aspect of comparative genomics. Orthologs in different species can be used to predict the function of uncontrived genes from model organisms as they retain the same biological function through the path of evolution. Orthologs can be inferred using phylogenetic, pair-wise similarity or synteny based methods. The study here describes a computational method for detecting orthologs of a protein. A phylogenetic tree based approach is used for identification of orthologous proteins. A Combination of species overlap algorithm and patristic distances is used for detecting orthologs of a protein from a set of FASTA sequences. Patristic distances have been used to drill the orthology predictions of any protein down to its closest orthologs. The approach gives a considerably good accuracy and has high specificity and precision. The use of Distance threshold allows controlling the stringency level of predictions so that the closeness and proximity between the protein of interest and its orthologs can be adjusted. Index Terms—Orthologs, comparative genomics, Phylogenetics, Species-overlap, Patristic distance.

A multiobjective memetic algorithm for PPI network alignment.

There recently has been great interest in aligning protein-protein interaction (PPI) networks to identify potentially orthologous proteins between species. It is thought that the topological information contained in these networks will yield better orthology predictions than sequence similarity alone. Recent work has found that existing aligners have difficulty making use of both topological and sequence similarity when aligning, with either one or the other being better matched. This can be at least partially attributed to the fact that existing aligners try to combine these two potentially conflicting objectives into a single objective. We present Optnetalign, a multiobjective memetic algorithm for the problem of PPI network alignment that uses extremely efficient swap-based local search, mutation and crossover operations to create a population of alignments. This algorithm optimizes the conflicting goals of topological and sequence similarity using the concept of Pareto dominance, exploring the tradeoff between the two objectives as it runs. This allows us to produce many high-quality candidate alignments in a single run. Our algorithm produces alignments that are much better compromises between topological and biological match quality than previous work, while better characterizing the diversity of possible good alignments between two networks. Our aligner's results have several interesting implications for future research on alignment evaluation, the design of network alignment objectives and the interpretation of alignment results. The C++ source code to our program, along with compilation and usage instructions, is available at https://github.com/crclark/optnetaligncpp/

Identification of divergent WH2 motifs by HMM-HMM alignments.

BackgroundThe actin cytoskeleton is a hallmark of eukaryotic cells. Its regulation as well as its interaction with other proteins is carefully orchestrated by actin interaction domains. One of the key players is the WH2 motif, which enables binding to actin monomers and filaments and is involved in the regulation of actin nucleation. Contrasting conserved domains, the identification of this motif in protein sequences is challenging, as it is short and poorly conserved.FindingsTo identify divergent members, we combined Hidden-Markov-Model (HMM) to HMM alignments with orthology predictions. Thereby, we identified nearly 500 proteins containing so far not annotated WH2 motifs. This included shootin-1, an actin binding protein involved in neuron polarization. Among others, WH2 motifs of ‘proximal to raf’ (ptr)-orthologs, which are described in the literature, but not annotated in genome databases, were identified.ConclusionIn summary, we increased the number of WH2 motif containing proteins substantially. This identification of candidate regions for actin interaction could steer their experimental characterization. Furthermore, the approach outlined here can easily be adapted to the identification of divergent members of further domain families.Electronic supplementary materialThe online version of this article (doi:10.1186/s13104-015-0981-7) contains supplementary material, which is available to authorized users.

Exploring the space of gene/species reconciliations with transfers.

Reconciliations between gene and species trees have important applications in the study of genome evolution (e.g. sequence orthology prediction or quantification of transfer events). While numerous methods have been proposed to infer them, little has been done to study the underlying reconciliation space. In this paper, we characterise the reconciliation space for two evolutionary models: the [Formula: see text] (duplication, loss and transfer) model and a variant of it-the no-[Formula: see text] model-which does not allow [Formula: see text] events (a transfer immediately followed by a loss). We provide formulae to compute the size of the corresponding spaces and define a set of transformation operators sufficient to explore the entire reconciliation space. We also define a distance between two reconciliations as the minimal number of operations needed to transform one into the other and prove that this distance is easily computable in the no-[Formula: see text] model. Computing this distance in the [Formula: see text] model is more difficult and it is an open question whether it is NP-hard or not. This work constitutes an important step toward reconciliation space characterisation and reconciliation comparison, needed to better assess the performance of reconciliation inference methods through simulations.

The OMA orthology database in 2015: function predictions, better plant support, synteny view and other improvements.

The Orthologous Matrix (OMA) project is a method and associated database inferring evolutionary relationships amongst currently 1706 complete proteomes (i.e. the protein sequence associated for every protein-coding gene in all genomes). In this update article, we present six major new developments in OMA: (i) a new web interface; (ii) Gene Ontology function predictions as part of the OMA pipeline; (iii) better support for plant genomes and in particular homeologs in the wheat genome; (iv) a new synteny viewer providing the genomic context of orthologs; (v) statically computed hierarchical orthologous groups subsets downloadable in OrthoXML format; and (vi) possibility to export parts of the all-against-all computations and to combine them with custom data for ‘client-side’ orthology prediction. OMA can be accessed through the OMA Browser and various programmatic interfaces at http://omabrowser.org.

A phylogeny-based benchmarking test for orthology inference reveals the limitations of function-based validation.

Accurate orthology prediction is crucial for many applications in the post-genomic era. The lack of broadly accepted benchmark tests precludes a comprehensive analysis of orthology inference. So far, functional annotation between orthologs serves as a performance proxy. However, this violates the fundamental principle of orthology as an evolutionary definition, while it is often not applicable due to limited experimental evidence for most species. Therefore, we constructed high quality "gold standard" orthologous groups that can serve as a benchmark set for orthology inference in bacterial species. Herein, we used this dataset to demonstrate 1) why a manually curated, phylogeny-based dataset is more appropriate for benchmarking orthology than other popular practices and 2) how it guides database design and parameterization through careful error quantification. More specifically, we illustrate how function-based tests often fail to identify false assignments, misjudging the true performance of orthology inference methods. We also examined how our dataset can instruct the selection of a “core” species repertoire to improve detection accuracy. We conclude that including more genomes at the proper evolutionary distances can influence the overall quality of orthology detection. The curated gene families, called Reference Orthologous Groups, are publicly available at http://eggnog.embl.de/orthobench2.

OrthoInspector 2.0: Software and database updates.

We previously developed OrthoInspector, a package incorporating an original algorithm for the detection of orthology and inparalogy relations between different species. We have added new functionalities to the package. While its original algorithm was not modified, performing similar orthology predictions, we facilitated the prediction of very large databases (thousands of proteomes), refurbished its graphical interface, added new visualization tools for comparative genomics/protein family analysis and facilitated its deployment in a network environment. Finally, we have released three online databases of precomputed orthology relationships. Package and databases are freely available at http://lbgi.fr/orthoinspector with all major browsers supported. odile.lecompte@unistra.fr Supplementary data are available at Bioinformatics online.

A transcriptomic approach to ribbon worm systematics (nemertea): resolving the pilidiophora problem.

Resolving the deep relationships of ancient animal lineages has proven difficult using standard Sanger-sequencing approaches with a handful of markers. We thus reassess the relatively well-studied phylogeny of the phylum Nemertea (ribbon worms)-for which the targeted gene approaches had resolved many clades but had left key phylogenetic gaps-by using a phylogenomic approach using Illumina-based de novo assembled transcriptomes and automatic orthology prediction methods. The analysis of a concatenated data set of 2,779 genes (411,138 amino acids) with about 78% gene occupancy and a reduced version with 95% gene occupancy, under evolutionary models accounting or not for site-specific amino acid replacement patterns results in a well-supported phylogeny that recovers all major accepted nemertean clades with the monophyly of Heteronemertea, Hoplonemertea, Monostilifera, being well supported. Significantly, all the ambiguous patterns inferred from Sanger-based approaches were resolved, namely the monophyly of Palaeonemertea and Pilidiophora. By testing for possible conflict in the analyzed supermatrix, we observed that concatenation was the best solution, and the results of the analyses should settle prior debates on nemertean phylogeny. The study highlights the importance, feasibility, and completeness of Illumina-based phylogenomic data matrices.

DOR - a Database of Olfactory Receptors - Integrated Repository for Sequence and Secondary Structural Information of Olfactory Receptors in Selected Eukaryotic Genomes.

Olfaction is the response to odors and is mediated by a class of membrane-bound proteins called olfactory receptors (ORs). An understanding of these receptors serves as a good model for basic signal transduction mechanisms and also provides important clues for the strategies adopted by organisms for their ultimate survival using chemosensory perception in search of food or defense against predators. Prior research on cross-genome phylogenetic analyses from our group motivated the addressal of conserved evolutionary trends, clustering, and ortholog prediction of ORs. The database of olfactory receptors (DOR) is a repository that provides sequence and structural information on ORs of selected organisms (such as Saccharomyces cerevisiae, Drosophila melanogaster, Caenorhabditis elegans, Mus musculus, and Homo sapiens). Users can download OR sequences, study predicted membrane topology, and obtain cross-genome sequence alignments and phylogeny, including three-dimensional (3D) structural models of 100 selected ORs and their predicted dimer interfaces. The database can be accessed from http://caps.ncbs.res.in/DOR. Such a database should be helpful in designing experiments on point mutations to probe into the possible dimerization modes of ORs and to even understand the evolutionary changes between different receptors.

Mutations in the UQCC1-Interacting Protein, UQCC2, Cause Human Complex III Deficiency Associated with Perturbed Cytochrome b Protein Expression

Mitochondrial oxidative phosphorylation (OXPHOS) is responsible for generating the majority of cellular ATP. Complex III (ubiquinol-cytochrome c oxidoreductase) is the third of five OXPHOS complexes. Complex III assembly relies on the coordinated expression of the mitochondrial and nuclear genomes, with 10 subunits encoded by nuclear DNA and one by mitochondrial DNA (mtDNA). Complex III deficiency is a debilitating and often fatal disorder that can arise from mutations in complex III subunit genes or one of three known complex III assembly factors. The molecular cause for complex III deficiency in about half of cases, however, is unknown and there are likely many complex III assembly factors yet to be identified. Here, we used Massively Parallel Sequencing to identify a homozygous splicing mutation in the gene encoding Ubiquinol-Cytochrome c Reductase Complex Assembly Factor 2 (UQCC2) in a consanguineous Lebanese patient displaying complex III deficiency, severe intrauterine growth retardation, neonatal lactic acidosis and renal tubular dysfunction. We prove causality of the mutation via lentiviral correction studies in patient fibroblasts. Sequence-profile based orthology prediction shows UQCC2 is an ortholog of the Saccharomyces cerevisiae complex III assembly factor, Cbp6p, although its sequence has diverged substantially. Co-purification studies show that UQCC2 interacts with UQCC1, the predicted ortholog of the Cbp6p binding partner, Cbp3p. Fibroblasts from the patient with UQCC2 mutations have deficiency of UQCC1, while UQCC1-depleted cells have reduced levels of UQCC2 and complex III. We show that UQCC1 binds the newly synthesized mtDNA-encoded cytochrome b subunit of complex III and that UQCC2 patient fibroblasts have specific defects in the synthesis or stability of cytochrome b. This work reveals a new cause for complex III deficiency that can assist future patient diagnosis, and provides insight into human complex III assembly by establishing that UQCC1 and UQCC2 are complex III assembly factors participating in cytochrome b biogenesis.