A Tool for Visualization and Analysis of Single-Cell RNA-Seq Data Based on Text Mining

Abstract

Gene expression in individual cells can now be measured for thousands of cells in a single experiment thanks to innovative sample-preparation and sequencing technologies. State-of-the-art computational pipelines for single-cell RNA-sequencing data, however, still employ computational methods that were developed for traditional bulk RNA-sequencing data, thus not accounting for the peculiarities of single-cell data, such as sparseness and zero-inflated counts. Here, we present a ready-to-use pipeline named gf-icf (gene frequency–inverse cell frequency) for normalization of raw counts, feature selection, and dimensionality reduction of scRNA-seq data for their visualization and subsequent analyses. Our work is based on a data transformation model named term frequency–inverse document frequency (TF-IDF), which has been extensively used in the field of text mining where extremely sparse and zero-inflated data are common. Using benchmark scRNA-seq datasets, we show that the gf-icf pipeline outperforms existing state-of-the-art methods in terms of improved visualization and ability to separate and distinguish different cell types.

Automated Extraction of Information From Texts of Scientific Publications: Insights Into HIV Treatment Strategies

“…The two case studies aimed at the extraction of chemical names from the texts relevant to HIV reverse transcriptase inhibition, proteins and genes from the texts relevant to HIV control allow us to determine the advantages and disadvantages of text mining approaches to new information. The main advantage of text mining approaches is the possibility of covering the huge amount of textual data (Ruusmann and Maran, 2013 ; Capuzzi et al, 2017 , 2018 ; Kandhro et al, 2017 ; Azam et al, 2019 ; Gambardella and di Bernardo, 2019 ; Guin et al, 2019 ; Ivanisenko et al, 2019 ; Alves et al, 2020 ). Text mining approaches allow retrieving the most recent and important information about chemicals, proteins, and genes associated with HIV treatment including their tissue-specific expression level (Ivanisenko et al, 2019 ).…”

A single-cell atlas of breast cancer cell lines to study tumour heterogeneity and drug response

“…We next generated an atlas encompassing all 32 cell-lines, as shown in Figure 1A and available online. This was achieved by combining data across cell-lines with the gf-icf pipeline ( 19 ), which performs count normalization, feature selection and dimensionality reduction ( 20 ) of the profiled cells. In the atlas, cell lines derived from the same cancer subtypes tend to cluster together, while being separated from the other subtypes (Figure 1A): luminal BC cell lines form a big “island” with multiple “peninsulas” with intermixing of cells from distinct cell lines (Figure 1A,B); on the contrary, TNBC cells give rise to an “archipelago” with cells from the same cell-line grouped into distinct islands, thus suggesting that TNBC cell-lines represent instances of distinct diseases.…”

“…Single cells expression profiles were normalized using GF-ICF (Gene Frequency – Inverse Cell Frequency) normalization using the gficf package 65,66 for R statistical environment (https://github.com/dibbelab/gficf). GF-ICF is based on a data transformation model called term frequency-inverse document frequency (TF-IDF) that has been extensively used in the field of text mining.…”

“…Cell clustering and identification of marker genes: Transcriptionally similar subpopulations of cells sux were found using a Phenograph like approach 67 as implemented in the clustcells function of gficf package 64 .…”

“…The Louvain algorithm with resolution svs parameter equal to 0.25 was used to find communities of cells in this graph. Differentially expressed genes svt in each cluster were identified by the findClusterMarkers function of gficf package, which compares the svu expression of a gene in each cluster versus all the other by using the Wilcoxon rank-sum test 64 .…”