Science Reviews - Biology, 2024, 3(2), 13-21 Martina Elena Tarozzi

Next Generation Sequencing Technologies, Bioinfor-

matics and Artificial Intelligence: A Shared Timeline

Martina Elena Tarozzi, PhD

Indipendent researcher. Florence, Tuscany, Italy

https://doi.org/10.57098/SciRevs.Biology.3.2.2

Received June 10, 2024. Revised June 25, 2024. Accepted June 26, 2024

Abstract: This review provides a comprehensive overview of the fast-paced and intertwined evolution of three

pivotal fields: next-generation sequencing (NGS) technologies, bioinformatics, and artificial intelligence (AI).

The paper begins by tracing the development of sequencing technologies and highlights how advancements in

genetic sequencing have led to an explosion of biological data, necessitating the rise of bioinformatics for data

management and analysis. The review next covers the primary steps and methods used in bioinformatic analysis

and concludes by reporting some of the technical and biological challenges in which AI methods have been

applied.

Introduction

Determining the order of nucleic acids in pol-

ynucleotide molecules and its functional meaning

has been a major biological question since the dis-

covery of the molecular structure of DNA in 1953

(1). Only twenty-four years later in 1977, the first

method for DNA sequencing was published by

Fredrick Sanger and colleagues, who developed the

“chain-termination" or dideoxy technique (2). Im-

provement of this groundbreaking method repre-

sented the first generation of DNA sequencing tech-

nologies, which produced reads nearly one kilobase

(kb) in length. The development of additional tech-

niques such as polymerase chain reaction (PCR) (3)

in 1983 and recombinant DNA technologies pro-

vided the means for generating high quantities of

DNA required by first-generation technologies,

triggering the genomic revolution and ultimately

the first draft of the human genome in 2001(4). A

pivotal turning point was achieved in 2005 with the

advent of Next Generation Sequencing (NGS) tech-

nologies (5), allowing for the massive and parallel

sequencing of whole genomes. In the last decade,

sequencing technologies have expanded to include

methods for RNA sequencing (6,7), giving rise to

the transcriptomic field and to methods for unveil-

ing the structural features and environmental-me-

diated modifications of chromatin and DNA (8), the

epigenomics field, and the single-cell omics technol-

ogies starting in 2009 (9).

Artificial intelligence (AI) was born and rap-

idly improved within this same time frame (Figure

1). Here again, the initial milestones date back to the

early 1950s, with the Turing test and the first use of

the term “Artificial Intelligence” at the Dartmouth

Conference by John McCarthy (10). In the following

three decades, the IT sector released groundbreak-

ing innovations, such as the Internet and World

Wide Web, as well as algorithms that laid the foun-

dation for deep learning, like Convolutional Neural

Networks (CNN), Support Vector Machine (SVM)

and Backpropagation (11–14). In less than a decade,

the AI Deep Blue was able to outperform human be-

ings in complex tasks such as playing chess. Starting

in 2001, the conjunction of increasingly powerful

computational resources (storage space and pro-

cessing speed) (15) and the biotechnological devel-

opment that led to NGS, increased the potential for

tandem use of AI and biology. NGS enabled a mas-

sive increase in omics data production, necessitat-

ing the development of computational methods

able to handle such data. The complexity of biolog-

ical processes and data provided opportunities and

challenges that machine learning techniques are

well suited to solve. Consequently, starting from

the early 2000s, computational biology became an

increasingly relevant field.

Martina Elena Tarozzi Science Reviews - Biology, 2024, 3(2), 13-21

As molecular biology becomes more data-in-

tensive and AI algorithms better able to handle bio-

logical complexity, the interconnection between

these fields is bound to strengthen. In this review,

we cover the main technological features of NGS

technologies and bioinformatic analysis and pro-

vide an overview of current applications of AI on

sequencing data.

Figure 1: Timeline of improvement milestones in genomics and sequencing technologies (blue arrows), informatics and artificial

intelligence (orange arrows) and computational biology (green arrows).

Overview of illumina technology and se-

quencing assays

NGS refers to modern high-throughput se-

quencing technologies that can be applied to DNA

or RNA. Illumina platforms are the NGS technology

most frequently used in research and clinical set-

tings, and will therefore be the focus in this paper.

The sequencing workflow starts with library

preparation, which varies between different omics.

In genomic workflows, DNA is fragmented either

mechanically, enzymatically, or with transposons in

fragments of appropriate length (typically around

400bp). Next, blunt ends at both ends are repaired:

typically, 5’ ends are phosphorylated and 3’ ends

are repaired with Adenine residues. Subsequently,

adapters are ligated to both ends. Depending on the

experimental design, it is then possible to select ge-

nome regions of interest via enrichment (such as the

exome or a more restricted selection of genes in tar-

get sequencing) or retain the entire genome.

RNA sequencing is a versatile high through-

put sequencing technique introduced in 2008 that

allows for the investigation of gene expression, as

well as alternative splicing , allele-specific expres-

sion , variation in linear nucleotide sequence, novel

transcript expression and gene fusion events . The

appropriate library preparation protocol for RNA

sequencing must be considered based on the

study’s questions in order to reduce bias . There are

four main categories of possible library prepara-

tions: i) total RNAseq: all structural, regulatory,

coding and non-coding RNAs are sequenced, ii)

RNAseq with ribosomal RNA reduction: only

rRNAs useful for phylogenic reconstruction are

kept together with regulatory and coding RNAs, iii)

cDNA capture: only coding RNAs are enriched us-

ing probes targeting exon sequences, iv) polyA se-

lection: only mature mRNAs are isolated through

poly-T probes that bind the 3’ poly-A tail of mRNAs.

The third main omic in terms of sequencing of

nucleic acids is epigenomics. Gene expression can

strongly depend on epigenetic regulation in terms

of changes in the accessibility of certain genomic re-

gions through DNA methylation, histone modifica-

tions, three-dimensional chromatin organization,

and post-transcriptional regulation mechanisms.

Each type of epigenetic modification can be meas-

Science Reviews - Biology, 2024, 3(2), 13-21 Martina Elena Tarozzi

ured via an experimental procedure performed be-

fore the library preparation described above. A

thorough overview of these complex techniques can

be found in Mehrmohamadi et al, 2021[24].

There are four main categories of epigenomic

assays: i) DNA methylation: normally investigated

with bisulfite-conversion-based libraries, where

only unmethylated cytosines are converted to uracil

(23) ii) histone modifications: mostly studied with

chromatin immunoprecipitation assays (ChIP-seq),

where crosslinked DNA is treated with antibodies

targeting the histone modification of interest and

pulled down (24), iii) chromatin accessibility: stud-

ied with assays based on the fact that open chroma-

tin is more accessible to fragmentation agents, like

digestion enzymes (e.g. DNase-seq) or transposase

(e.g. ATAC-seq)(25), iv) 3D organization: mostly in-

vestigated with Chromatin Conformation Capture

(3C) derived assays (such as 4C-seq, Hi-C, ChIA-

Drop), whereby nuclei DNA is crosslinked, chi-

meric DNA molecules made of genomic regions

close to one another are formed, and the proximal

genomic regions in the 3D space of the nucleus are

measured using pairwise frequencies between ge-

nomic loci (26).

Once the library is loaded into the sequencer,

it is pumped onto a flow-cell where each single-

stranded fragment hybridizes to flow-cell adapters

on both ends forming a “bridge” structure that con-

fers the name to this step called “bridge amplifica-

tion” (27). After cluster generation, sequencing by

synthesis with reversible chain terminators and a

fluorophore corresponding to each of the four nu-

cleotides (A, C, T, and G), or bases, takes place. The

fluorophore wavelength together with its intensity

determines the base call. The acquired optical sig-

nals for each lane of the flow cell are converted into

base call format files (bcl file) that represent the out-

put of sequencing and the first raw input for bioin-

formatic analysis (Figure 2).

Figure 2: Schematic representation of a targeted genomic library preparation workflow and Illumina sequencing reaction. Figure

adapted from images courtesy of www.illumina.com (https://www.illumina.com/science/technology/next-generation-sequenc-

ing.html, https://www.illumina.com/documents/products/techspotlights/techspotlight_sequencing.pdf ).

Bioinformatic analysis

Bioinformatic analysis involves several steps

of computationally intensive data transformations,

each requiring specific tools. Substantial research

has been done to develop reliable software capable

of performing such data transformations and to

make the computational analysis of NGS data re-

producible. For example, package and environment

management systems such as CONDA allow the in-

stallation and management of software developed

in different languages for different operating sys-

tems.

Martina Elena Tarozzi Science Reviews - Biology, 2024, 3(2), 13-21

3.1. Primary Data Analysis

The primary analysis of Illumina raw data

consists in demultiplexing bcl data into FASTQ files.

This process transforms the binary raw files ob-

tained from the optic signal acquired during se-

quencing into a text file in which the nucleotide se-

quence is matched by a PHRED score defining the

quality of the nucleotide, or base, call made by the

sequencer. During this step, index sequences con-

tained in each fragment are used to group all calls

belonging to the same sample into a single FASTQ

file. Depending on the sequencer, this step is per-

formed directly on-instrument by Real-Time Anal-

ysis software or afterwards as first step of a com-

mand line pipeline (28).

3.2 Secondary Data Analysis

Secondary analysis of genomic data is typi-

cally a standardized workflow used in diagnostics

and research . It generally consists of three main

steps: mapping onto the reference genome, post-

alignment processing and, if useful, variant calling.

Workflow management systems such as Snake-

make and Nextflow address the need to concate-

nate the steps required by bioinformatic pipelines

while providing an efficient usage of computational

power through process parallelization. The sam-

ple’s sequence is expressed in FASTQ files contain-

ing the nucleotide sequence and quality scores de-

rived from millions of short reads. Quality checks

are typically performed first in secondary analysis.

One of the most common tools for this task is

FASTQC . Adapter reads and base calls with low

quality scores are trimmed from the sequence using

specific tools such as Trimmomatic , fastp or cu-

tadapt . Mapping algorithms are then used to iden-

tify a location in the reference genome that matches

the experimental read generated via sequencing.

Software settings allow for varying degrees of toler-

ance towards base mismatches and extra spaces to

allow for the detection of possible variants. One

commonly used tool for mapping NGS sequences to

a reference genome is the Burrows-Wheeler Align-

ment (BWA). The output of mapping algorithms are

stored as sam (sequence alignment map) files,

which contain all the information surrounding the

mapping procedure in the metadata section, along

with data concerning the mapped genomic region

and mapping. The binary counterpart to a sam file

is a bam file, which represents the type of data that

will undergo further post-alignment processing

and variant calling. Post-alignment processing con-

sists of sorting, marking duplicates and indexing

the bam file. The variant calling step aims to iden-

tify single nucleotide variants and small insertions

and deletions, reported in Variant Call Format (VCF)

files.

3.3 Tertiary Data Analysis

Tertiary analysis of genomic data uses the list

of variants reported in the VCF file to biologically

interpret the data. This component of data analysis

is highly adaptable depending on the experimental

question. The first and most common step is variant

annotation, aimed at obtaining functional infor-

mation about the type of nucleotide substitution

and its predicted effect. A more classic approach is

to consider only the variants that affect the primary

structure of the coded protein, like missense or

truncating variants, especially if in-silico variant

predictor tools describe them as likely pathogenic.

Functional analysis on groups of significant genes is

often performed with network approaches or over-

representation methods. In recent years, data sci-

ence and machine learning methods applied to ge-

nomic data have been a resourceful approach for ac-

quiring a more complete understanding of poly-

genic contributions in complex diseases (39) and in

the context of precision medicine (40,41). Some of

these applications will be further discussed section

5. Applications of ai in genomics and transcriptomics

ML and DL methods have been applied to se-

quencing data covering various research scopes and

topics. Here, we describe some of the most relevant

fields of application, focusing first on the use of ML

and DL for technical issues associated with NGS

data processing and analysis, and then with exam-

ples of how these methods are used to explore open

biological questions. This section aims at providing

common, promising or exemplifying applications

of AI methods on NGS data in biology and bioinfor-

matics and should not be considered a complete

overview of all its possible applications.

5.1 Applications on technical problems

Secondary bioinformatic analysis: variant calling

Science Reviews - Biology, 2024, 3(2), 13-21 Martina Elena Tarozzi

The accurate discovery of single nucleotide

variants from billions of short reads remains a chal-

lenging step in bioinformatics because library prep-

aration, sequencing and data processing tools are

error-prone procedures. These issues become even

more apparent when the object of study are low-fre-

quency somatic mutations or when the input DNA

is of lower quality. Most variant callers use statisti-

cal methods (such as logistic regression, hidden

Markov models, naïve Bayes) to model error

sources and to distinguish whether differences be-

tween experimental reads and the reference ge-

nome are caused by true genetic variants or errors.

In recent years, deep learning has been applied to

address variant calling on NGS data: a common ap-

proach is to address the problem as one of image

recognition, where a Deep Neural Network ana-

lyzes sequencing data that are transformed as im-

ages of read pileups of true genotype calls to com-

pute the genotype likelihoods at each locus. Two of

the first and most popular tools of this kind are

DeepVariant (42) and DeNovoCNN (43), with the

latter specifically used to address the identification

of de novo mutations. Both tools showed higher ac-

curacy compared to classical methods. An alterna-

tive approach is presented in HELLO (44), whereby

comparable performances are obtained by design-

ing Deep Neural Networks that examine aligned

reads to predict the status (ref or alt) of each candi-

date allele given the support for that allele in rela-

tion to the support for the remaining alleles at the

genomic site.

Tertiary bioinformatic analysis: Variant effect prediction

Variant Effect Prediction (VEP) are computa-

tional tools that provide a prediction about the func-

tional significance of a single nucleotide variant

(SNV). The growing use of NGS technologies for

advanced diagnostics has increased the need to bet-

ter classify variants of uncertain significance. VEPs

rely on different types of prior knowledge, such as

protein sequence and structural information, evolu-

tionary sequence conservation, functional experi-

ments, epigenomic data and association studies to

produce an effect score for the variant. In super-

vised VEPs, the algorithm is trained on a set of la-

belled SNVs known as benign or dam2aging ac-

cording to previous knowledge to perform a classi-

fication task. Using this prior knowledge, these

methods compute a score expressing the predicted

effect of the variant. Examples of well-performing

supervised VEPs on human samples are SNP&GO

(45), PolyPhen2 (46) and DEOGEN2 (47). Unsuper-

vised methods do not use any labelled data and

usually rely exclusively on the evolutionary conser-

vation of the genomic locus. This group also in-

cludes deep learning methods, like DeepSequence

(48), considered by a recent benchmark study as the

top-performing tool among 46 tested in deep muta-

tion scanning data (49). An example of a semi-su-

pervised deep learning method is the Illumina Pri-

mateAI (50), which has performed well in the study

of rare diseases.

Visualization of high dimensional datasets

NGS data are highly dimensional because

each sample is sequenced simultaneously. The huge

amount of information contained in these data can

represent an obstacle to the identification of its most

meaningful features. Dimensionality reduction

techniques such as PCA, t-SNE and UMAP are used

to identify latent components in the data that are

not easily accessible due to the high number of var-

iables. Data are thus transformed into a lower di-

mensionality while maintaining the relationships

between data points (e.g., samples) as much as pos-

sible. These methods are extremely versatile. For ex-

ample, they can be used in the pre-processing of

bulk RNA-seq data to identify possible outliers and

relevant covariates(51), to search for recurrent pat-

terns on targeted DNA sequencing data in different

classes of samples (52), or to visualize single-cell

RNA sequencing data. In this context, dimensional-

ity reduction techniques coupled with clustering al-

gorithms are used for cell-type identification tasks,

identifying groups of cells that share similar expres-

sion profiles. Another application of these methods

is lineage trajectory inference(53), which involves

the reconstruction of the position of each individual

cell on the lineage trajectory based on scRNA-seq

profiles with different time points, allowing for the

study of dynamic processes such as the cell cycle,

cell differentiation and cell activation.

6. Future Perspectives

In this review, we summarized the crucial as-

pects and timeline of NGS technologies, bioinfor-

matics and AI, highlighted how they are connected

in a holistic process, and explained the potential

revolutionary insights that can be gained from their

Martina Elena Tarozzi Science Reviews - Biology, 2024, 3(2), 13-21

concurrent use. Computational and molecular biol-

ogy have and are continuing to advance at an im-

pressive pace. Machine and deep learning, while

relatively recent breakthroughs in the biological

and biomedical fields, will almost certainly play an

increasing role. Substantial investments are being

made by leading technology companies that, to-

gether with academic researchers, are implement-

ing innovative methodologies, software and archi-

tectures tailored specifically to answer biological

questions. Concurrently, we are witnessing the im-

provement of molecular techniques while sequenc-

ing experiments are evolving and becoming more

affordable, as evident by the growing interest in

long-read sequencing. Taken together, these inno-

vations are a new frontier of research and have the

potential to strongly affect our ability understand

and interact with genetic information.

References

1. Watson JD, Crick FHC. Molecular Structure of Nucleic Acids: A Structure for Deoxyribose Nucleic

Acid. Nat 1953 1714356 [Internet]. 1953 Apr 25 [cited 2021 Jul 6];171(4356):737–8. Available from:

https://www.nature.com/articles/171737a0

2. Sanger F, Nicklen S, Coulson AR. DNA sequencing with chain-terminating inhibitors. Proc Natl

Acad Sci U S A [Internet]. 1977 [cited 2021 Jul 6];74(12):5463. Available from:

/pmc/articles/PMC431765/?report=abstract

3. Mullis K, Faloona F, Scharf S, Saiki R, Horn G, Erlich H. Specific enzymatic amplification of DNA

in vitro: The polymerase chain reaction. Cold Spring Harb Symp Quant Biol. 1986;51(1):263–73.

4. JC V, MD A, EW M, PW L, RJ M, GG S, et al. The sequence of the human genome. Science

[Internet]. 2001 Feb 16 [cited 2021 Jul 6];291(5507):1304–51. Available from:

https://pubmed.ncbi.nlm.nih.gov/11181995/

5. Heather J, Chain B, Heather JM, Chain B. The Sequence of Sequencers : The History of Sequencing

DNA Genomics The sequence of sequencers : The history of sequencing DNA. Genomics [Internet].

2015;107(1):1–8. Available from: http://dx.doi.org/10.1016/j.ygeno.2015.11.003

6. Nagalakshmi U, Wang Z, Waern K, Shou C, Raha D, Gerstein M, et al. The Transcriptional

Landscape of the Yeast Genome Defined by RNA Sequencing. Science [Internet]. 2008 Jun 6 [cited

2021 Jul 19];320(5881):1344. Available from: /pmc/articles/PMC2951732/

7. SJ E, WB B, L L, PS S. Gene discovery and annotation using LCM-454 transcriptome sequencing.

Genome Res [Internet]. 2007 Jan [cited 2021 Jul 6];17(1):69–73. Available from:

https://pubmed.ncbi.nlm.nih.gov/17095711/

8. JFriedman N, Rando OJ. Epigenomics and the structure of the living genome. Genome Res.

2015;25(10):1482–90.

9. Wang D, Bodovitz S. Single cell analysis: the new frontier in ‘Omics.’ Trends Biotechnol [Internet].

2010 Jun [cited 2021 Jul 6];28(6):281. Available from: /pmc/articles/PMC2876223/

10. McCarthy J, Minsky ML, Rochester N, Shannon CE. A proposal for the Dartmouth summer

research project on artificial intelligence [Internet]. Vol. 27, AI Magazine. 2006 [cited 2022 Oct 12]. p.

12–4. Available from: http://www-formal.stanford.edu/jmc/history/dartmouth/dartmouth.html

11. LeCun Y, Boser B, Denker JS, Henderson D, Howard RE, Hubbard W, et al. Backpropagation

applied to digit recognition [Internet]. Vol. 1, Neural computation. 1989. p. 541–51. Available from:

https://www.ics.uci.edu/~welling/teaching/273ASpring09/lecun-89e.pdf

12. Lecun Y, Bottou E, Bengio Y, Haffner P. Gradient-Based Learning Applied to Document

Recognition. 1998.

13. Rosenblatt F. The perceptron: A probabilistic model for information storage and organization in

the brain. Psychol Rev. 1958;65(6):386–408.

Science Reviews - Biology, 2024, 3(2), 13-21 Martina Elena Tarozzi

14. Boser BE, Guyon IM, Vapnik VN. Training algorithm for optimal margin classifiers. Proc Fifth

Annu ACM Work Comput Learn Theory. 1992;144–52.

15. Muir P, Li S, Lou S, Wang D, Spakowicz DJ, Salichos L, et al. The real cost of sequencing: Scaling

computation to keep pace with data generation. Genome Biol [Internet]. 2016 Mar 23 [cited 2023 Apr

28];17(1):1–9. Available from: https://genomebiology.biomedcentral.com/articles/10.1186/s13059-

016-0917-0

16. Lister R, O’Malley RC, Tonti-Filippini J, Gregory BD, Berry CC, Millar AH, et al. Highly

Integrated Single-Base Resolution Maps of the Epigenome in Arabidopsis. Cell [Internet]. 2008 May

2 [cited 2021 Jul 19];133(3):523. Available from: /pmc/articles/PMC2723732/

17. U N, Z W, K W, C S, D R, M G, et al. The transcriptional landscape of the yeast genome defined

by RNA sequencing. Science [Internet]. 2008 Jun 6 [cited 2021 Jul 6];320(5881):1344–9. Available from:

https://pubmed.ncbi.nlm.nih.gov/18451266/

18. Mehmood A, Laiho A, Venäläinen MS, McGlinchey AJ, Wang N, Elo LL. Systematic evaluation

of differential splicing tools for RNA-seq studies. Brief Bioinform [Internet]. 2020 Dec 1 [cited 2021

Jul 20];21(6):2052–65. Available from: https://academic.oup.com/bib/article/21/6/2052/5648232

19. Castel SE, Levy-Moonshine A, Mohammadi P, Banks E, Lappalainen T. Tools and best practices

for data processing in allelic expression analysis. Genome Biol [Internet]. 2015;16(1):1–13. Available

from: http://dx.doi.org/10.1186/s13059-015-0762-6

20. Hutchins AP, Poulain S, Fujii H, Miranda-Saavedra D. Discovery and characterization of new

transcripts from RNA-seq data in mouse CD4+ T cells. Genomics. 2012 Nov 1;100(5):303–13.

21. Haas BJ, Dobin A, Stransky N, Li B, Yang X, Tickle T, et al. STAR-Fusion: Fast and Accurate Fusion

Transcript Detection from RNA-Seq. bioRxiv [Internet]. 2017 Mar 24 [cited 2021 Jul 20];120295.

Available from: https://www.biorxiv.org/content/10.1101/120295v1

22. Van Dijk EL, Jaszczyszyn Y, Thermes C. Library preparation methods for next-generation

sequencing: Tone down the bias. Exp Cell Res. 2014 Mar 10;322(1):12–20.

23. Mehrmohamadi M, Sepehri MH, Nazer N, Norouzi MR. A Comparative Overview of Epigenomic

Profiling Methods. Front Cell Dev Biol. 2021;9(July):1–14.

24. Landt SG, Marinov GK, Kundaje A, Kheradpour P, Pauli F, Batzoglou S, et al. ChIP-seq guidelines

and practices of the ENCODE and modENCODE consortia. Genome Res [Internet]. 2012 Sep 1 [cited

2022 Oct 27];22(9):1813–31. Available from: https://genome.cshlp.org/content/22/9/1813.full

25. Buenrostro JD, Wu B, Chang HY, Greenleaf WJ. ATAC-seq: A Method for Assaying Chromatin

Accessibility Genome-Wide. Curr Protoc Mol Biol [Internet]. 2015 [cited 2022 Oct 27];109:21.29.1.

Available from: /pmc/articles/PMC4374986/

26. Lieberman-Aiden E, Van Berkum NL, Williams L, Imakaev M, Ragoczy T, Telling A, et al.

Comprehensive mapping of long-range interactions reveals folding principles of the human

genome. Science (80- ) [Internet]. 2009 Oct 9 [cited 2021 Jul 5];326(5950):289–93. Available from:

/pmc/articles/PMC2858594

27. Bentley DR, Balasubramanian S, Swerdlow HP, Smith GP, Milton J, Brown CG, et al. Accurate

Whole Human Genome Sequencing using Reversible Terminator Chemistry. Nature [Internet]. 2008

Nov 11 [cited 2022 Oct 18];456(7218):53. Available from: /pmc/articles/PMC2581791/

28. Richards S, Aziz N, Bale S, Bick D, Das S, Gastier-Foster J, et al. Standards and guidelines for the

interpretation of sequence variants: A joint consensus recommendation of the American College of

Medical Genetics and Genomics and the Association for Molecular Pathology. Genet Med

[Internet]. 2015 May 8 [cited 2021 Mar 31]; 17(5):405–24. Available from:

/pmc/articles/PMC4544753/

Martina Elena Tarozzi Science Reviews - Biology, 2024, 3(2), 13-21

29. Alonso CM, Llop M, Sargas C, Pedrola L, Panadero J, Hervás D, et al. Clinical Utility of a Next-

Generation Sequencing Panel for Acute Myeloid Leukemia Diagnostics. J Mol Diagn [Internet]. 2019

Mar 1 [cited 2022 Oct 13];21(2):228–40. Available from: https://pubmed.ncbi.nlm.nih.gov/30576870/

30. Di Resta C, Galbiati S, Carrera P, Ferrari M. Next-generation sequencing approach for the

diagnosis of human diseases: open challenges and new opportunities. EJIFCC [Internet]. 2018 Apr 1

[cited 2022 Oct 13];29(1):4. Available from: /pmc/articles/PMC5949614/

31. Bartoletti-Stella A, Tarozzi M, Mengozzi G, Asirelli F, Brancaleoni L, Mometto N, et al. Dementia-

related genetic variants in an Italian population of early-onset Alzheimer’s disease. Front Aging

Neurosci. 2022;14(September):1–13.

32. Köster J, Rahmann S. Snakemake-a scalable bioinformatics workflow engine. Bioinformatics.

2012.

33. DI Tommaso P, Chatzou M, Floden EW, Barja PP, Palumbo E, Notredame C. Nextflow enables

reproducible computational workflows. Nat Biotechnol 2017 354 [Internet]. 2017 Apr 11 [cited 2023

Mar 2];35(4):316–9. Available from: https://www.nature.com/articles/nbt.3820

34. Andrews, Simon, Krueger, Felix , Segonds-Pichon, Anne , Biggins, Laura , Krueger, Christel ,

Wingett S. FastQC [Internet]. Babraham, UK. 2010. Available from:

http://www.bioinformatics.babraham.ac.uk/projects/fastqc

35. Bolger AM, Lohse M, Usadel B. Trimmomatic: a flexible trimmer for Illumina sequence data.

Bioinformatics [Internet]. 2014 Aug 1 [cited 2021 Nov 17];30(15):2114–20. Available from:

https://academic.oup.com/bioinformatics/article/30/15/2114/2390096 metanalysis of machine

and deep learning-based CRISPR gRNA design tools. RNA Biol. 2020; 17:13

36. Chen S, Zhou Y, Chen Y, Gu J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics

[Internet]. 2018 Sep 1 [cited 2023 Jan 10];34(17):i884–90. Available from:

https://academic.oup.com/bioinformatics/article/34/17/i884/5093234

37. Martin M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMB-

net.journal [Internet]. 2011 May 2 [cited 2023 Jan 10];17(1):10–2. Available from: https://journal.emb-

net.org/index.php/embnetjournal/article/view/200/479

38. Li H, Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioin-

formatics. 2009;

39. Ho DSW, Schierding W, Wake M, Saffery R, O’Sullivan J. Machine learning SNP based prediction

for precision medicine. Front Genet. 2019;10(MAR):267.

40. Dias R, Torkamani A. Artificial intelligence in clinical and genomic diagnostics. Genome Med.

2019;11(1):1–12.

41. Álvarez-Machancoses Ó, Galiana EJD, Cernea A, de la Viña JF, Fernández-Martínez JL. On the

role of artificial intelligence in genomics to enhance precision medicine. Pharmgenomics Pers Med.

2020;13:105–19.

42. Poplin R, Chang PC, Alexander D, Schwartz S, Colthurst T, Ku A, et al. A universal snp and small-

indel variant caller using deep neural networks. Nat Biotechnol. 2018;36(10):983.

43. Khazeeva G, Sablauskas K, van der Sanden B, Steyaert W, Kwint M, Rots D, et al. DeNovoCNN:

a deep learning approach to de novo variant calling in next generation sequencing data. Nucleic Ac-

ids Res. 2022;50(17):e97.

44. Ramachandran A, Lumetta SS, Klee EW, Chen D. HELLO: improved neural network architectures

and methodologies for small variant calling. BMC Bioinformatics [Internet]. 2021;22(1):1–31. Availa-

ble from: https://doi.org/10.1186/s12859-021-04311-4

Science Reviews - Biology, 2024, 3(2), 13-21 Martina Elena Tarozzi

45. Capriotti E, Calabrese R, Fariselli P, Martelli PL, Altman RB, Casadio R. WS-SNPs&GO: a web

server for predicting the deleterious effect of human protein variants using functional annotation.

BMC Genomics [Internet]. 2013;14 Suppl 3(Suppl 3):S6. Available from: http://www.biomedcen-

tral.com/1471-2164/14/S3/S6

46. Adzhubei I, Jordan DM, Sunyaev SR. Predicting Functional Effect of Human Missense Mutations

Using PolyPhen-2. Curr Protoc Hum Genet [Internet]. 2013 [cited 2021

47. Raimondi D, Tanyalcin I, FertCrossed JSD, Gazzo A, Orlando G, Lenaerts T, et al. DEOGEN2:

Prediction and interactive visualization of single amino acid variant deleteriousness in human pro-

teins. Nucleic Acids Res. 2017;45(W1):W201–6.

48. Riesselman AJ, Ingraham JB, Marks DS. Deep generative models of genetic variation capture the

effects of mutations. Nat Methods [Internet]. 2018;15(10):816–22. Available from:

http://dx.doi.org/10.1038/s41592-018-0138-4

49. Livesey BJ, Marsh JA. Using deep mutational scanning to benchmark variant effect predictors and

identify disease mutations. Mol Syst Biol [Internet]. 2020 Jul 1 [cited 2022 Oct 26];16(7):e9380. Availa-

ble from: https://onlinelibrary.wiley.com/doi/full/10.15252/msb.20199380

50. Sundaram L, Gao H, Padigepati SR, McRae JF, Li Y, Kosmicki JA, et al. Predicting the clinical

impact of human mutation with deep neural networks. Nat Genet 2018 508 [Internet]. 2018 Jul 23

[cited 2022 Oct 27];50(8):1161–70. Available from: https://www.nature.com/articles/s41588-018-

0167-z

51. Conesa A, Madrigal P, Tarazona S, Gomez-Cabrero D, Cervera A, McPherson A, et al. A survey

of best practices for RNA-seq data analysis. Genome Biol. 2016;17(1):1–19.

52. Tarozzi M, Bartoletti-Stella A, Dall’Olio D, Matteuzzi T, Baiardi S, Parchi P, et al. Identification of

recurrent genetic patterns from targeted sequencing panels with advanced data science: a case-study

on sporadic and genetic neurodegenerative diseases. BMC Med Genomics 2022 151 [Internet]. 2022

Feb 10 [cited 2022 Feb 25];15(1):1–12. Available from: https://bmcmedgenomics.biomedcen-

tral.com/articles/10.1186/s12920-022-01173-4

53. Saelens W, Cannoodt R, Todorov H, Saeys Y. A comparison of single-cell trajectory inference

methods. Nat Biotechnol [Internet]. 2019;37(5):547–54. Available from:

http://dx.doi.org/10.1038/s41587-019-0071-9