8

There are two types of features which can be extracted from protein sequences. It is not possible to known which, if any, of these features would be useful in classification tasks. It may not be possible to build a classifier at all, or it may be very straight forward. To know this, a feature selection technique must be used, such as e.g. forward feature ...


7

R's nnet package only supports fully connected neural networks with one hidden layer. This is the most primitive type of network. I doubt it will work well for promotors finding. In addition, such network won't give you useful interpretations. If you want to explore neural networks, you should use a one-dimension convolution layer as is described in this ...


5

TL;DR: docking is much slower than any ML approach, but the ML approach can be constrained by pharmacophores dictated by the active site. Side note: Scale The scale for ligand space exploration is generally is generally several orders of magnitude higher than "hundreds": Zinc DB lists 750 million enumerated compounds, GDB13 has enumerate all possible ...


5

While your question is specific to cancerous germline mutations, I'd suggest you look at the COSMIC database of somatic mutations to include in your analysis. There are other factors to include in this kind of analysis you're suggesting, such as predictive deleterious effects (PolyPhen for example can perform such predictions). If you have 10M variants/...


4

Short answer: you can't. Neural networks use positive and negative examples to add weights to the neural network architecture that is provided to it. Trying to deconvolve the meaning behind the weights is nigh-on impossible except for the simplest perceptron. This is common to many machine learning algorithms: they work very much like black boxes. One ...


3

I think MEME is a good tool for your purpose, but there are others as well. I can think of InterProScan for example (although I am not sure if it is really de novo). Here a summary of available tools.


3

I don't think it will be possible to do what you ask, right now with current knowledge. Selecting variants relevant to cancer risk is still an open problem and usually requires quite a lot of human intervention. You can use different measures of population frequency to filter common variants with the assumption that frequent variants won't be pathogenic. ...


2

I am not sure if I understood correctly how you classify the drugs. But what you attempt to do is similar to what they do here but in that article they use the drugs targeted to a molecule/pathway to find combination of drugs that are better for the disease (not an alternative drug), see the first image: In that article they mention the pathway cross-talk ...


2

One option is to treat all proteins that are not explicitly known to form a biofilm as proteins that are incapable of doing so, but this will likely result in many false negatives. You could take a look at one-class classification and positive-unlabeled (PU) learning. Both are techniques specifically designed for these kind of problems where only one of the ...


2

Since you are using R, you probably don't want to use scikit-learn, which is for Python. However, there is a similar R library mlr ("R package to make machine learning in R easy") that provides a unified interface to all popular machine learning methods. Check their tutorial on how to get started.


2

Kaggle has a cats v dogs dataset if you want to try your hand at biological image classification: https://www.kaggle.com/c/dogs-vs-cats The DREAM challenge competitions are usually specifically oriented to solving biological classification or learning problems with ML: http://dreamchallenges.org/


2

I believe you're referring to something related to multi-layer networks. Here you can find an interesting reading about this kind of networks: https://academic.oup.com/comnet/article/2/3/203/2841130 Also you can try with this one: https://www.sciencedirect.com/science/article/pii/S0370157314002105 As a matter of fact, I'm not an expert of multi-layer ...


2

I come from a protein background and this problem is analogous to protein torsion angle prediction, which in turn is a variant of protein secondary structure prediction. Conventional ways to go here would be to use a sliding window of x bases either side to make a prediction for each point, or use a CNN/LSTM-type model to run over the whole sequence and ...


2

The question of folding of RNA sequences looks slightly similar to protein folding - perhaps searching in this domain might bring more suggestions. An example of (current) state of the art of deep learning approach to protein folding is AlphaFold developed by DeepMind for CASP competition . They basically decompose the angles to an image representation and ...


2

Responded in the plink2-users group (https://groups.google.com/forum/#!topic/plink2-users/zgJxdXxvdLo ).


2

The UMAP results look very pretty :) If I were to try to make sense of this blob, I'd try to do the following. Pull out Gene Ontology (GO) terms for a list of candidate genes (you can try to subset Uniprot / Swissprot for some examples). Then I feed through some genes through your fragmented BERT kmer approach to see where they land. I'd be particularly ...


2

The steps you describe are correct. For step 2 it is usually normalized to mean 0 and variance 1. However the "machine learning" part is important. Having several samples being technical replicates will make the integration task easier. However, you have too few samples to make any good prediction. At most I would describe it as an exploratory analysis. I ...


2

Use Fst (migration statistic) such as the Fstat program to generate a distance matrix and solve it via clustering, such as UPGMA or neighbor-joining. This analysis is particularly useful for diploids. I presume you have heterozygotes in your analysis. You can switch Fis (inbreeding) if you have excessive homozygosity. The other way in is a haplotype network. ...


2

Not to discount @Michael's answer (unsupervised learning is very handy for descriptive analysis of this kind of problem), but classification based on microsatellite data should not be too difficult. For papers with examples, you can see e.g. this and this. They seem to apply various algorithms, you might be able to learn something from those papers. More ...


2

You could try the PBMC 3k data from the Satija lab: https://satijalab.org/seurat/articles/pbmc3k_tutorial.html For well-annotated data, there's the Single-cell proteo-genomic reference map: https://www.biorxiv.org/content/10.1101/2021.03.18.435922v1 Data for that are available here. If these datasets are too large (i.e. too good), you can use the existing ...


1

Actually the paper has made it clear how they did it. You just have to read the supplementary materials closer. In their figure from the paper: A is Red, C is Green, G is blue, and T is Yellow (G+R), but this is still unclear how they the 3xNxN image. In RGB, each dimension is an NxN image. Since you have three dimensions, so it's 3xNxN. The red dimension ...


1

This is a comment to redirect to other answers, but go too long. Docking is the in silico prediction of where a ligand binds in a protein. There are several Q&As about docking here that are relevant, so please check out: machine learning, docking and the factors that affect it: Generate ligands candidates based on protein shape —also worth a read the &...


1

I see a lot of people using xCell. There should also be papers systematically compare many different methods, which might interest you.


1

The concern with your approach is called 'leakage' because you are parameterising the same data set that you are training. This can easily lead to overtraining is in ANN (artificial neural networks) and is not cool. Overtraining is where the ANN very tightly wraps around the training data set that it performs poorly on any other data set. In other words it ...


1

In truth deep neural networks quite mainstream now, albeit they are a pain because of their reliance on very expensive graphics cards to process the data (GPU) rather than traditional CPU. There are two issues, The algorhithms need an application to a biological phenomena and this is dependent on the model. DNA applications (usually) use a different model ...


1

This answer is no longer on topic. The question is about what is the best machine learning algorithm to use to analyse atomic coordinates from an MD trajectory in order to infer a novel properties.   Also, it should be preface for any confused reader that unfortunately "feature" is a technical term in both genetics and statistics. For the former a ...


1

I understand that Phenotype is a set of criteria that you apply on your EHR data to select patients of interest from EHR database. No, a phenotype is a way of behavior or other observed characteristic of a person, resulting from a combination on its underlying biology and the environment it's in. Think things like, height, running speed, strength, etc. In ...


1

I did some research and according to Wikipedia an "enrichment factor" in terms of bioinformatics is... Where possible, it's best to use explanations provided in the original source. This is because definitions could be different, depending on who is writing. Here's some descriptive information from the paper that may be more informative: The enrichment ...


1

I have been using dada2 and phyloseq (phangorn for trees) combination to analyze a V4 region 16s rRNA data set. I find their approach and explanations very understandable and usable since I prefer R. I was under the impression that even full-length 16s rRNA gene sequences were inadequate to classify up to species level since closely related species within a ...


1

If you are training a machine learning algorithm you probably want to train against things you know to be true, rather than making predictions from the sequence and training against the predictions. In this case you will want to use protein structures with the relevant ligands bound as your data. If you go to the RCSB PDB advanced search you can search for ...


Only top voted, non community-wiki answers of a minimum length are eligible