Leukemia Prediction Using Sparse Logistic Regression

We describe a supervised prediction method for diagnosis of acute myeloid leukemia (AML) from patient samples based on flow cytometry measurements. We use a data driven approach with machine learning methods to train a computational model that takes in flow cytometry measurements from a single patient and gives a confidence score of the patient being AML-positive. Our solution is based on an regularized logistic regression model that aggregates AML test statistics calculated from individual test tubes with different cell populations and fluorescent markers. The model construction is entirely data driven and no prior biological knowledge is used. The described solution scored a 100% classification accuracy in the DREAM6/FlowCAP2 Molecular Classification of Acute Myeloid Leukaemia Challenge against a golden standard consisting of 20 AML-positive and 160 healthy patients. Here we perform a more extensive validation of the prediction model performance and further improve and simplify our original method showing that statistically equal results can be obtained by using simple average marker intensities as features in the logistic regression model. In addition to the logistic regression based model, we also present other classification models and compare their performance quantitatively. The key benefit in our prediction method compared to other solutions with similar performance is that our model only uses a small fraction of the flow cytometry measurements making our solution highly economical.     

Molecular dynamics simulations of a branched tetradecasaccharide substrate in the active site of a xyloglucan endo-transglycosylase

Molecular dynamics simulations of the tetradecasaccharide XXXGXXXG in complex with the hybrid aspen xyloglucan endo-transglycosylase PttXET16-34 have been performed and analysed with respect to structure, dynamics, flexibility and ligand interactions. Notably, the charge state of the so-called ‘helper residue’ aspartate 87 (Asp87), which lies between the catalytic nucleophile [glutamate 85 (Glu85)] and general acid/base (Glu89) residues on the same beta strand, had a significant effect on PttXET16-34 active site structure. When Asp87 was deprotonated, electrostatic repulsion forced the nucleophile away from C1 of the sugar ring in subsite ? 1 and the proton–donating ability of Glu89 was also weakened due to the formation of a hydrogen bond with Asp87, whereas the protonation of Asp87 resulted in the formation of a hydrogen bond with the catalytic nucleophile and correct positioning of the catalytic machinery. The results suggest that catalysis in glycoside hydrolase family 16, and by extension clan GH-B enzymes, is optimal when the catalytic nucleophile is deprotonated for nucleophilic attack on the substrate, whereas the ‘helper residue’ and general acid/base residue are both in their conjugate acid forms to align the nucleophile and deliver a proton to the departing sugar, respectively.     

Mammalian and Malaria Parasite Cyclase-associated Proteins Catalyze Nucleotide Exchange on G-actin through a Conserved Mechanism

Cyclase-associated proteins (CAPs) are among the most highly conserved regulators of actin dynamics, being present in organisms from mammals to apicomplexan parasites. Yeast, plant, and mammalian CAPs are large multidomain proteins, which catalyze nucleotide exchange on actin monomers from ADP to ATP and recycle actin monomers from actin-depolymerizing factor (ADF)/cofilin for new rounds of filament assembly. However, the mechanism by which CAPs promote nucleotide exchange is not known. Furthermore, how apicomplexan CAPs, which lack many domains present in yeast and mammalian CAPs, contribute to actin dynamics is not understood. We show that, like yeast Srv2/CAP, mouse CAP1 interacts with ADF/cofilin and ADP-G-actin through its N-terminal α-helical and C-terminal β-strand domains, respectively. However, in the variation to yeast Srv2/CAP, mouse CAP1 has two adjacent profilin-binding sites, and it interacts with ATP-actin monomers with high affinity through its WH2 domain. Importantly, we revealed that the C-terminal β-sheet domain of mouse CAP1 is essential and sufficient for catalyzing nucleotide exchange on actin monomers, although the adjacent WH2 domain is not required for this function. Supporting these data, we show that the malaria parasite Plasmodium falciparum CAP, which is entirely composed of the β-sheet domain, efficiently promotes nucleotide exchange on actin monomers. Collectively, this study provides evidence that catalyzing nucleotide exchange on actin monomers via the β-sheet domain is the most highly conserved function of CAPs from mammals to apicomplexan parasites. Other functions, including interactions with profilin and ADF/cofilin, evolved in more complex organisms to adjust the specific role of CAPs in actin dynamics.     

Separating water‐potential induced swelling and shrinking from measured radial stem variations reveals a cambial growth and osmotic concentration signal

The quantification of cambial growth over short time periods has been hampered by problems to discern between growth and the swelling and shrinking of a tree stem. This paper presents a model, which separates cambial growth and reversible water‐potential induced diurnal changes from simultaneously measured whole stem and xylem radial variations, from field‐measured Scots pine trees in Finland. The modelled growth, which includes osmotic concentration changes, was compared with (direct) dendrometer measurements and microcore samples. In addition, the relationship of modelled growth and dendrometer measurements to environmental factors was analysed. The results showed that the water‐potential induced changes of tree radius were successfully separated from stem growth. Daily growth predicted by the model exhibited a high correlation with the modelled daily changes of osmotic concentration in phloem, and a temperature dependency in early summer. Late‐summer growth saw higher dependency on water availability and temperature. Evaluation of the model against dendrometer measurements showed that the latter masked a true environmental signal in stem growth due to water‐potential induced changes. The model provides better understanding of radial growth physiology and offers potential to examine growth dynamics and changes due to osmotic concentration, and how the environment affects growth.     

New separation-free assay technique for SNPs using two-photon excitation fluorometry

A new separation-free method for detection of single nucleotide polymorphisms (SNPs) is described. The method is based on the single base extension principle, fluorescently labeled dideoxy nucleotides and two-photon fluorescence excitation technology, known as ArcDia™ TPX technology. In this assay technique, template-directed single base extension is carried out for primers which have been immobilized on polymer microparticles. Depending on the sequence of the template DNA, the primers are extended either with a labeled or with a non-labeled nucleotide. The genotype of the sample is determined on the basis of two-photon excited fluorescence of individual microparticles. The effect of various assay condition parameters on the performance of the assay method is studied. The performance of the new assay method is demonstrated by genotyping the SNPs of human individuals using double-stranded PCR amplicons as samples. The results show that the new SNP assay method provides sensitivity and reliability comparable to the state-of-the-art SNaPshot™ assay method. Applicability of the new method in routine laboratory use is discussed with respect to alternative assay techniques.     

The dibenzodioxocin lignin substructure is abundant in the inner part of the secondary wall in Norway spruce and silver birch xylem

A specific condensed lignin substructure, dibenzodioxocin, was immunolocalized in differentiating cell walls of Norway spruce (Picea abies (L.) H. Karsten) and silver birch (Betula pendula Roth) xylem. A fluorescent probe, Alexa 488 was used as a marker on the dibenzodioxocin-specific secondary antibody. For the detection of this lignin substructure, 25-m cross-sections of xylem were viewed with a confocal laser-scanning microscope with fluorescein isothiocyanate fluorescence filters. In mature cells, fluorescence was detected in the S3 layer of the secondary wall in both tree species, but it was more intense in Norway spruce than in silver birch. In silver birch most of the signal was detected in vessel walls and less in fiber cell walls. In very young tracheids of Norway spruce and vessels and fibers of silver birch, where secondary cell wall layers were not yet formed, the presence of the dibenzodioxocin structure could not be shown.Abbreviation CLSM confocal laser-scanning fluorescence microscopy     

Biological role and structural mechanism of twinfilin-capping protein interaction

Twinfilin and capping protein (CP) are highly conserved actin-binding proteins that regulate cytoskeletal dynamics in organisms from yeast to mammals. Twinfilin binds actin monomer, while CP binds the barbed end of the actin filament. Remarkably, twinfilin and CP also bind directly to each other, but the mechanism and role of this interaction in actin dynamics are not defined. Here, we found that the binding of twinfilin to CP does not affect the binding of either protein to actin. Furthermore, site-directed mutagenesis studies revealed that the CP-binding site resides in the conserved C-terminal tail region of twinfilin. The solution structure of the twinfilin-CP complex supports these conclusions. In vivo, twinfilin's binding to both CP and actin monomer was found to be necessary for twinfilin's role in actin assembly dynamics, based on genetic studies with mutants that have defined biochemical functions. Our results support a novel model for how sequential interactions between actin monomers, twinfilin, CP, and actin filaments promote cytoskeletal dynamics.     

BAPS 2: enhanced possibilities for the analysis of genetic population structure

Bayesian statistical methods based on simulation techniques have recently been shown to provide powerful tools for the analysis of genetic population structure. We have previously developed a Markov chain Monte Carlo (MCMC) algorithm for characterizing genetically divergent groups based on molecular markers and geographical sampling design of the dataset. However, for large-scale datasets such algorithms may get stuck to local maxima in the parameter space. Therefore, we have modified our earlier algorithm to support multiple parallel MCMC chains, with enhanced features that enable considerably faster and more reliable estimation compared to the earlier version of the algorithm. We consider also a hierarchical tree representation, from which a Bayesian model-averaged structure estimate can be extracted. The algorithm is implemented in a computer program that features a user-friendly interface and built-in graphics. The enhanced features are illustrated by analyses of simulated data and an extensive human molecular dataset. AVAILABILITY: Freely available at http://www.rni.helsinki.fi/~jic/bapspage.html.