Software review: the KNIME workflow environment and its applications in genetic programming and machine learning

O’Hagan, Steve; Kell, Douglas B.

doi:10.1007/s10710-015-9247-3

Cited by 40 publications

(37 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The structure of the interrogating molecule, provided in the form of a SMILES [240] or InChI [241][242][243] string, is encoded therefrom as a string of 1s and 0s. Modern software suites such as RDKit [244] (www.rdkit.org/), CDK [245] or KNIME (which includes them both) [246][247][248] allow one to do this automatically. The same is done for the molecule(s) with which the interrogating molecule is to be compared.…”

Section: The Importance Of Qsars (Quantitative Structure-activity Relmentioning

confidence: 99%

“…The result of a series of such measurements using various drugs is a set of paired values of structure (encoded as a bitstring of 1s and 0s as described above) and the value of the objective function. Any number of the modern methods of machine learning may then be used to construct a nonlinear mapping of these inputs onto the output; in [248] we compared three, viz a multivariate linear statistical method known as partial least squares [257], a potent general method known as random forests [258], and an evolutionary algorithm known as genetic programming (see e.g. [259][260][261][262]).…”

Section: The Importance Of Qsars (Quantitative Structure-activity Relmentioning

confidence: 99%

See 1 more Smart Citation

Control of metabolite efflux in microbial cell factories: current advances and future prospects

Kell¹

2018

Preprint

View full text Add to dashboard Cite

The yield and ease of purification of biotechnological products are typically greatly enhanced if they can be secreted into the external medium. Although not universally appreciated, however, small molecule biotechnological products do not ‘float across’ the phospholipid bilayer portion of biological membranes, and thus they need transporters to assist their passage into the extramembrane and extracellular spaces. Some of these transporters may be reversible, equilibrative transporters that might more normally be used for uptake, while others may have an efflux directionality imposed on them via suitable energy coupling mechanisms. Despite the energetic costs of small molecule synthesis, natural evolution does in fact provide a number of mechanisms by which the secretion of such products can actually enhance fitness. Where available these provide useful starting points, and assaying for such activities is crucial. In particular, in a systems biology approach, we first need to identify such/suitable activities before we can seek to increase them. They may then be improved through promoter engineering, via manipulation of control elements, or by directed evolution of the transporter proteins themselves. Modern methods of synthetic biology provide enormous opportunities for all kinds of efflux transporter engineering; they are just beginning to be realised.

show abstract

Section: The Importance Of Qsars (Quantitative Structure-activity Relmentioning

confidence: 99%

Section: The Importance Of Qsars (Quantitative Structure-activity Relmentioning

confidence: 99%

Control of metabolite efflux in microbial cell factories: current advances and future prospects

Kell¹

2018

Preprint

View full text Add to dashboard Cite

show abstract

“…The experiments were performed with use of The Konstanz Information Miner environment (KNIME), Version 3 [18], [19].…”

Section: Resultsmentioning

confidence: 99%

Preprocessing compensation techniques for improved classification of imbalanced medical datasets

Wosiak

Karbowiak

2017

Annals of Computer Science and Information Systems

View full text Add to dashboard Cite

Abstract-The paper describes the study on the problem of applying classification techniques in medical datasets with a class imbalance. The aim of the research is to identify factors that negatively affect classification results and propose actions that may be taken to improve the performance. To alleviate the impact of uneven and complex class distribution, methods of balancing the datasets are proposed and compared. The experiments were conducted on five datasets -three binary and two multiclass. They comprise several data preprocessing methods applied on data and the classification with different techniques. The study shows that for some datasets there exists a combination of a certain preprocessing method and a classification technique which outperforms other approaches. For datasets with complex distribution or too many features the ratio of correctly predicted labels may be low regardless what resampling method and classification technique has been applied.

show abstract

“…These were all performed as in our previous work of this type (O'Hagan and Kell, 2015a,b,c; O'Hagan et al, 2015; O'Hagan and Kell, 2016), using the KNIME workflow system (see e.g., Berthold et al, 2008; Mazanetz et al, 2012; Meinl et al, 2012; Warr, 2012; O'Hagan and Kell, 2015b and http://knime.org/). …”

Section: Methodsmentioning

confidence: 99%

Enhancing Drug Efficacy and Therapeutic Index through Cheminformatics-Based Selection of Small Molecule Binary Weapons That Improve Transporter-Mediated Targeting: A Cytotoxicity System Based on Gemcitabine

et al. 2017

Self Cite

View full text Add to dashboard Cite

The transport of drug molecules is mainly determined by the distribution of influx and efflux transporters for which they are substrates. To enable tissue targeting, we sought to develop the idea that we might affect the transporter-mediated disposition of small-molecule drugs via the addition of a second small molecule that of itself had no inhibitory pharmacological effect but that influenced the expression of transporters for the primary drug. We refer to this as a “binary weapon” strategy. The experimental system tested the ability of a molecule that on its own had no cytotoxic effect to increase the toxicity of the nucleoside analog gemcitabine to Panc1 pancreatic cancer cells. An initial phenotypic screen of a 500-member polar drug (fragment) library yielded three “hits.” The structures of 20 of the other 2,000 members of this library suite had a Tanimoto similarity greater than 0.7 to those of the initial hits, and each was itself a hit (the cheminformatics thus providing for a massive enrichment). We chose the top six representatives for further study. They fell into three clusters whose members bore reasonable structural similarities to each other (two were in fact isomers), lending strength to the self-consistency of both our conceptual and experimental strategies. Existing literature had suggested that indole-3-carbinol might play a similar role to that of our fragments, but in our hands it was without effect; nor was it structurally similar to any of our hits. As there was no evidence that the fragments could affect toxicity directly, we looked for effects on transporter transcript levels. In our hands, only the ENT1-3 uptake and ABCC2,3,4,5, and 10 efflux transporters displayed measurable transcripts in Panc1 cultures, along with a ribonucleoside reductase RRM1 known to affect gemcitabine toxicity. Very strikingly, the addition of gemcitabine alone increased the expression of the transcript for ABCC2 (MRP2) by more than 12-fold, and that of RRM1 by more than fourfold, and each of the fragment “hits” served to reverse this. However, an inhibitor of ABCC2 was without significant effect, implying that RRM1 was possibly the more significant player. These effects were somewhat selective for Panc cells. It seems, therefore, that while the effects we measured were here mediated more by efflux than influx transporters, and potentially by other means, the binary weapon idea is hereby fully confirmed: it is indeed possible to find molecules that manipulate the expression of transporters that are involved in the bioactivity of a pharmaceutical drug. This opens up an entirely new area, that of chemical genomics-based drug targeting.

show abstract

Software review: the KNIME workflow environment and its applications in genetic programming and machine learning

Cited by 40 publications

References 7 publications

Control of metabolite efflux in microbial cell factories: current advances and future prospects

Control of metabolite efflux in microbial cell factories: current advances and future prospects

Preprocessing compensation techniques for improved classification of imbalanced medical datasets

Enhancing Drug Efficacy and Therapeutic Index through Cheminformatics-Based Selection of Small Molecule Binary Weapons That Improve Transporter-Mediated Targeting: A Cytotoxicity System Based on Gemcitabine

Contact Info

Product

Resources

About