Table 3.

Number of occurences of most common learning algorithms in a database of PubMed publications of the last ten years on Next Generation Sequencing, single cell sequencing, gene expression and transcriptomics. Algorithms occurring less frequently have been omitted. The PubMed query was performed on 16th January 2021. The table also reports the specific biological applications of such methods in the selected database (although it does not represent an exhaustive list in general). Explicit bibliographic references can be found on our Github page and in the Supplementary Materials together with more details on the methodology used to build the table.

Learning algorithm	Example biological applications	#Occurrences
Support Vector Machine	diagnostic classification, intratumoral heterogeneity; tissue-selective genes; gene prediction; gene selection; disease-gene association; gene expression analysis; signatures from gene-pathway; disease gene prioritization; miRNA signatures	157
Random Forest	diagnostic classification, tissue-selective genes; gene prediction; co-acting gene networks; gene selection; mutation-gene-drug relations; gene expression analysis; miRNA biomarkers; drug-induced gene expression; sample-classification	124
Logistic Regression	gene prediction; gene selection; drug-induced gene expression	38
Deep Neural Network	mutation-gene-drug relations; gene expression analysis	36
LASSO	intratumoral heterogeneity; biomarkers selection; gene selection; gene expression analysis	31
Naive Bayes	gene selection; pharmacogenetic prediction	28
K-Nearest Neighbor	gene selection	28
Artificial Neural Network	gene selection; genotype-phenotype analysis; risk classification; transcriptome profiling; variant extraction	25
Autoencoder	gene prediction	24
Principal Component Analysis	single-cell analysis; gene expression	19
Linear Discriminant Analysis	transcriptome profiling; miRNA biomarkers; taxa-condition association	14
Perceptron	gene selection; gene prediction	12
K-means	candidate miRNA targets	8