Non-Negative Matrix Factorization for Learning Alignment-Specific Models of Protein Evolution

Abstract

Models of protein evolution currently come in two flavors: generalist and specialist. Generalist models (e.g. PAM, JTT, WAG) adopt a one-size-fits-all approach, where a single model is estimated from a number of different protein alignments. Specialist models (e.g. mtREV, rtREV, HIVbetween) can be estimated when a large quantity of data are available for a single organism or gene, and are intended for use on that organism or gene only. Unsurprisingly, specialist models outperform generalist models, but in most instances there simply are not enough data available to estimate them. We propose a method for estimating alignment-specific models of protein evolution in which the complexity of the model is adapted to suit the richness of the data. Our method uses non-negative matrix factorization (NNMF) to learn a set of basis matrices from a general dataset containing a large number of alignments of different proteins, thus capturing the dimensions of important variation. It then learns a set of weights that are specific to the organism or gene of interest and for which only a smaller dataset is available. Thus the alignment-specific model is obtained as a weighted sum of the basis matrices. Having been constrained to vary along only as many dimensions as the data justify, the model has far fewer parameters than would be required to estimate a specialist model. We show that our NNMF procedure produces models that outperform existing methods on all but one of 50 test alignments. The basis matrices we obtain confirm the expectation that amino acid properties tend to be conserved, and allow us to quantify, on specific alignments, how the strength of conservation varies across different properties. We also apply our new models to phylogeny inference and show that the resulting phylogenies are different from, and have improved likelihood over, those inferred under standard models.

Restricted Hidden Cardinality Constraints in Causal Models

“…The problem of performing a nonnegative matrix factorization is not always tractable, nevertheless numerical approximations of this task exist. This method is used in many fields of science, such as astronomy, bioinformatics and machine learning, as they study data with positive entries [3,4,17]. Furthermore, the nonnegative rank, which we define below, plays a critical role in communication theory as it characterizes randomized communication complexity and randomized correlation complexity [14,27].…”

A comprehensive genealogy of the replication associated protein of CRESS DNA viruses reveals a single origin of intron-containing Rep

“…Furthermore, since most CRESS DNA viruses have very few ORFs, often only a CP and a Rep, their sequence diversity contributes little to the overall Pfam database. While general models do increasingly well at describing the patterns dominated by genetic code constraints and the physiochemical properties of the amino acids (Murrell et al, 2011), there will always be opportunities for specific matrices to resolve protein evolution in biological entities that have more unique lifestyles and constraints, such as fast-evolving CRESS DNA viruses with single-stranded mutational biases (Cardinale et al, 2013;Frederico et al, 1990;Xia and Yuen, 2005).…”

Revealing Kinetics of Paracetamol Crystallization Using Time-Resolved Raman Spectroscopy, Orthogonal Time-Lapse Photography, and Non-Negative Matrix Factorization (OSANO)

“…Acoustic levitator allows for studying in situ phase transformations because of its contact-less nature and ability to minimize external influences. − The problem of finding the time-dependent concentration of intermediates of the nucleation process is equivalent to finding non-negative matrix factors of raw data of time-resolved Raman spectroscopy. Non-negative matrix factorization (NMF) is a linear algebraic method that finds applications in various scientific disciplines such as in the correction of astronomical spectroscopy, document clustering, and bioinformatics for latent component detection. − The two resulting factors of NMF provide qualitative and quantitative contribution of a single feature (latent components) to a complex phenomenon. In order to compute these contributions, Weber et al developed a robust clustering method known as PCCA+.…”