View source for Machine learning in bioinformatics

{{Use mdy dates|date=August 2022}}
{{Artificial intelligence}}
'''Machine learning in bioinformatics''' is the application of [[machine learning]] algorithms to [[bioinformatics]],<ref>{{cite journal | vauthors = Chicco D | title = Ten quick tips for machine learning in computational biology | journal = BioData Mining | volume = 10 | issue = 35 | pages = 35 | date = December 2017 | pmid = 29234465 | pmc = 5721660 | doi = 10.1186/s13040-017-0155-3 | doi-access = free }}</ref> including [[genomics]], [[proteomics]], [[microarrays]], [[systems biology]], [[evolution]], and [[text mining]].<ref name="Larrañaga2006">{{cite journal|display-authors=6|vauthors=Larrañaga P, Calvo B, Santana R, Bielza C, Galdiano J, Inza I, Lozano JA, Armañanzas R, Santafé G, Pérez A, Robles V|date=March 2006|title=Machine learning in bioinformatics|journal=Briefings in Bioinformatics|volume=7|issue=1|pages=86–112|doi=10.1093/bib/bbk007|pmid=16761367|doi-access=free}}</ref><ref name="Pérez-Wohlfeil2018">{{cite journal|vauthors=Pérez-Wohlfeil E, Torrenoa O, Bellis LJ, Fernandes PL, Leskosek B, Trellesa O|date=December 2018|title=Training bioinformaticians in High Performance Computing|journal=Heliyon|volume=4|issue=12|pages=e01057|doi=10.1016/j.heliyon.2018.e01057|doi-access=free |pmc=6299036|pmid=30582061|bibcode=2018Heliy...401057P }}</ref>

Prior to the emergence of machine learning, bioinformatics algorithms had to be programmed by hand; for problems such as [[protein structure prediction]], this proved difficult.<ref name=":2">{{cite journal | vauthors = Yang Y, Gao J, Wang J, Heffernan R, Hanson J, Paliwal K, Zhou Y | title = Sixty-five years of the long march in protein secondary structure prediction: the final stretch? | journal = Briefings in Bioinformatics | volume = 19 | issue = 3 | pages = 482–494 | date = May 2018 | pmid = 28040746 | pmc = 5952956 | doi = 10.1093/bib/bbw129 }}</ref> Machine learning techniques such as [[deep learning]] can [[feature learning|learn features]] of data sets rather than requiring the programmer to define them individually. The algorithm can further learn how to combine low-level [[Feature (machine learning)|features]] into more abstract features, and so on. This multi-layered approach allows such systems to make sophisticated predictions when appropriately trained. These methods contrast with other [[computational biology]] approaches which, while exploiting existing datasets, do not allow the data to be interpreted and analyzed in unanticipated ways. 

{{Toclimit|3}}

== Tasks ==
Machine learning algorithms in bioinformatics can be used for prediction, classification, and feature selection. Methods to achieve this task are varied and span many disciplines; most well known among them are machine learning and statistics. Classification and prediction tasks aim at building models that describe and distinguish classes or concepts for future prediction. The differences between them are the following:

* Classification/recognition outputs a categorical class, while prediction outputs a numerical valued feature.
* The type of algorithm, or process used to build the predictive models from data using analogies, rules, neural networks, probabilities, and/or statistics.

Due to the exponential growth of information technologies and applicable models, including artificial intelligence and data mining, in addition to the access ever-more comprehensive data sets, new and better information analysis techniques have been created, based on their ability to learn. Such models allow reach beyond description and provide insights in the form of testable models.

== Machine learning approaches ==

=== Artificial neural networks===
[[Artificial neural network]]s in bioinformatics have been used for:<ref>{{cite book|chapter-url=https://link.springer.com/chapter/10.1007%2F978-981-15-2445-5_3|title=Statistical Modelling and Machine Learning Principles for Bioinformatics Techniques, Tools, and Applications|vauthors=Shastry KA, Sanjay HA|date=2020|publisher=Springer|isbn=978-981-15-2445-5|veditors=Srinivasa K, Siddesh G, Manisekhar S|series=Algorithms for Intelligent Systems|location=Singapore|pages=25–39|chapter=Machine Learning for Bioinformatics|doi=10.1007/978-981-15-2445-5_3|s2cid=214350490|access-date=June 28, 2021}}</ref>

* Comparing and aligning RNA, protein, and DNA sequences.
* Identification of promoters and finding genes from sequences related to DNA.
* Interpreting the expression-gene and micro-array data.
* Identifying the network (regulatory) of genes.
* Learning evolutionary relationships by constructing [[phylogenetic tree]]s.
* Classifying and predicting [[protein structure]].
* [[Molecular modelling|Molecular design]] and [[Docking (molecular)|docking]]

=== Feature engineering ===
The way that features, often vectors in a many-dimensional space, are extracted from the domain data is an important component of learning systems.<ref name=":0">{{cite journal |vauthors= Soueidan H, Nikolski M|date=2019|title= Machine learning for metagenomics: methods and tools|issn=2449-7657|journal=Metagenomics|volume= 1|doi= 10.1515/metgen-2016-0001 |arxiv=1510.06621|s2cid=17418188}}</ref> In genomics, a typical representation of a sequence is a vector of [[k-mer]]s frequencies, which is a vector of dimension <math>4^k</math> whose entries count the appearance of each subsequence of length <math>k</math> in a given sequence. Since for a value as small as <math>k=12</math> the dimensionality of these vectors is huge (e.g. in this case the dimension is <math>4^{12} \approx 16\times 10^6</math>), techniques such as [[principal component analysis]] are used to project the data to a lower dimensional space, thus selecting a smaller set of features from the sequences.<ref name=":0" />{{Additional citation needed|date=June 2021}}

=== Classification ===

In this type of machine learning task, the output is a discrete variable. One example of this type of task in bioinformatics is labeling new genomic data (such as genomes of unculturable bacteria) based on a model of already labeled data.<ref name=":0" />

==== Hidden Markov models ====
[[Hidden Markov models]] (HMMs) are a class of [[statistical models]] for sequential data (often related to systems evolving over time). An HMM is composed of two mathematical objects: an observed state‐dependent process <math>X_1,X_2,\ldots ,X_M</math>, and an unobserved (hidden) state process <math>S_1,S_2,\ldots ,S_T</math>. In an HMM, the state process is not directly observed – it is a 'hidden' (or 'latent') variable – but observations are made of a state‐dependent process (or observation process) that is driven by the underlying state process (and which can thus be regarded as a noisy measurement of the system states of interest).<ref>{{Cite journal| vauthors = Rabiner L, Juang B |date=January 1986|title=An introduction to hidden Markov models|url=https://ieeexplore.ieee.org/document/1165342|journal=IEEE ASSP Magazine|volume=3|issue=1|pages=4–16|doi=10.1109/MASSP.1986.1165342|s2cid=11358505|issn=1558-1284}}</ref> HMMs can be formulated in continuous time.<ref>{{cite journal|vauthors=Jackson CH, Sharples LD, Thompson SG, Duffy SW, Couto E|date=July 2003|title=Multistate Markov models for disease progression with classification error|journal=Journal of the Royal Statistical Society, Series D (The Statistician)|volume=52|issue=2|pages=193–209|doi=10.1111/1467-9884.00351|s2cid=9824404 }}</ref><ref>{{cite journal|vauthors=Amoros R, King R, Toyoda H, Kumada T, Johnson PJ, Bird TG|date=May 30, 2019|title=A continuous-time hidden Markov model for cancer surveillance using serum biomarkers with application to hepatocellular carcinoma|journal=Metron|volume=77|issue=2|pages=67–86|doi=10.1007/s40300-019-00151-8|pmc=6820468|pmid=31708595}}</ref>

HMMs can be used to profile and convert a multiple sequence alignment into a position-specific scoring system suitable for searching databases for homologous sequences remotely.<ref>{{cite journal | vauthors = Eddy SR | title = Profile hidden Markov models | journal = Bioinformatics | volume = 14 | issue = 9 | pages = 755–63 | date = October 1, 1998 | pmid = 9918945 | doi = 10.1093/bioinformatics/14.9.755 | doi-access = free }}</ref> Additionally, ecological phenomena can be described by HMMs.<ref>{{cite journal | vauthors = McClintock BT, Langrock R, Gimenez O, Cam E, Borchers DL, Glennie R, Patterson TA | title = Uncovering ecological state dynamics with hidden Markov models | journal = Ecology Letters | volume = 23 | issue = 12 | pages = 1878–1903 | date = December 2020 | pmid = 33073921 | doi = 10.1111/ele.13610 | pmc = 7702077 | arxiv = 2002.10497 | bibcode = 2020EcolL..23.1878M }}</ref>

==== Convolutional neural networks ====
[[Convolutional neural network]]s (CNN) are a class of [[deep neural network]] whose architecture is based on shared weights of convolution kernels or filters that slide along input features, providing translation-equivariant responses known as feature maps.<ref name=":CNN-0">{{cite journal| vauthors = Zhang W |date=1988|title=Shift-invariant pattern recognition neural network and its optical architecture|url=https://drive.google.com/file/d/1nN_5odSG_QVae54EsQN_qSz-0ZsX6wA0/view?usp=sharing|journal=Proceedings of Annual Conference of the Japan Society of Applied Physics}}</ref><ref name=":CNN-1">{{cite journal | vauthors = Zhang W, Itoh K, Tanida J, Ichioka Y | title = Parallel distributed processing model with local space-invariant interconnections and its optical architecture | journal = Applied Optics | volume = 29 | issue = 32 | pages = 4790–7 | date = November 1990 | pmid = 20577468 | doi = 10.1364/AO.29.004790 | bibcode = 1990ApOpt..29.4790Z }}</ref> CNNs take advantage of the hierarchical pattern in data and assemble patterns of increasing complexity using smaller and simpler patterns discovered via their filters.<ref name=":7">{{Cite book |last=Bishop |first=Christopher M. |title=Pattern Recognition and Machine Learning |date=2006-08-17 |publisher=Springer |isbn=978-0-387-31073-2 |location=New York |language=English}}</ref>

Convolutional networks were [[mathematical biology|inspired]] by [[biological]] processes<ref name=fukuneoscholar>{{cite journal | vauthors = Fukushima K |year=2007 |title=Neocognitron |journal=Scholarpedia |volume=2 |issue=1| page=1717 |doi=10.4249/scholarpedia.1717| bibcode=2007SchpJ...2.1717F |doi-access=free }}</ref><ref name="hubelwiesel1968">{{cite journal | vauthors = Hubel DH, Wiesel TN | title = Receptive fields and functional architecture of monkey striate cortex | journal = The Journal of Physiology | volume = 195 | issue = 1 | pages = 215–43 | date = March 1968 | pmid = 4966457 | pmc = 1557912 | doi = 10.1113/jphysiol.1968.sp008455 }}</ref><ref name="intro">{{cite journal | vauthors = Fukushima K | title = Neocognitron: a self organizing neural network model for a mechanism of pattern recognition unaffected by shift in position | journal = Biological Cybernetics | volume = 36 | issue = 4 | pages = 193–202 | year = 1980 | pmid = 7370364 | doi = 10.1007/BF00344251 | s2cid = 206775608 }}</ref><ref name="robust face detection">{{cite journal | vauthors = Matsugu M, Mori K, Mitari Y, Kaneda Y | title = Subject independent facial expression recognition with robust face detection using a convolutional neural network | journal = Neural Networks | volume = 16 | issue = 5–6 | pages = 555–9 | year = 2003 | pmid = 12850007 | doi = 10.1016/S0893-6080(03)00115-1 }}</ref> in that the connectivity pattern between [[Artificial neuron|neurons]] resembles the organization of the animal [[visual cortex]]. Individual [[cortical neuron]]s respond to stimuli only in a restricted region of the [[visual field]] known as the [[receptive field]]. The receptive fields of different neurons partially overlap such that they cover the entire visual field.

CNN uses relatively little pre-processing compared to other [[Image classification|image classification algorithms]]. This means that the network learns to optimize the [[Filter (signal processing)|filters]] (or kernels) through automated learning, whereas in traditional algorithms these filters are [[Feature engineering|hand-engineered]]. This reduced reliance on prior knowledge of the analyst and on human intervention in manual feature extraction makes CNNs a desirable model.<ref name=":7" />

A phylogenetic convolutional neural network (Ph-CNN) is a [[convolutional neural network]] architecture proposed by Fioranti et al. in 2018 to classify [[metagenomics]] data.<ref name=":6">{{cite journal | vauthors = Fioravanti D, Giarratano Y, Maggio V, Agostinelli C, Chierici M, Jurman G, Furlanello C | title = Phylogenetic convolutional neural networks in metagenomics | journal = BMC Bioinformatics | volume = 19 | issue = Suppl 2 | pages = 49 | date = March 2018 | pmid = 29536822 | doi = 10.1186/s12859-018-2033-5 | pmc = 5850953 | doi-access = free }}</ref> In this approach, phylogenetic data is endowed with patristic distance (the sum of the lengths of all branches connecting two [[operational taxonomic unit]]s [OTU]) to select k-neighborhoods for each OTU, and each OTU and its neighbors are processed with convolutional filters.

==== Self-supervised learning ====
Unlike supervised methods, [[self-supervised learning]] methods learn representations without relying on annotated data. That is well-suited for genomics, where [[high throughput sequencing]] techniques can create potentially large amounts of unlabeled data. Some examples of self-supervised learning methods applied on genomics include DNABERT and Self-GenomeNet.<ref>{{Cite journal |last1=Ji |first1=Yanrong |last2=Zhou |first2=Zhihan |last3=Liu |first3=Han |last4=Davuluri |first4=Ramana V |date=2021-08-09 |editor-last=Kelso |editor-first=Janet |title=DNABERT: pre-trained Bidirectional Encoder Representations from Transformers model for DNA-language in genome |url=https://academic.oup.com/bioinformatics/article/37/15/2112/6128680 |journal=Bioinformatics |language=en |volume=37 |issue=15 |pages=2112–2120 |doi=10.1093/bioinformatics/btab083 |pmid=33538820 |issn=1367-4803|pmc=11025658 }}</ref><ref>{{Cite journal |last1=Gündüz |first1=Hüseyin Anil |last2=Binder |first2=Martin |last3=To |first3=Xiao-Yin |last4=Mreches |first4=René |last5=Bischl |first5=Bernd |last6=McHardy |first6=Alice C. |last7=Münch |first7=Philipp C. |last8=Rezaei |first8=Mina |date=2023-09-11 |title=A self-supervised deep learning method for data-efficient training in genomics |journal=Communications Biology |language=en |volume=6 |issue=1 |page=928 |doi=10.1038/s42003-023-05310-2 |pmid=37696966 |issn=2399-3642|doi-access=free |pmc=10495322 }}</ref>

==== Random forest ====
[[File: Some bioinformatic applications of Random Forest.jpg|thumbnail|Some bioinformatic applications{{Which|date=June 2021}} of Random Forest|220x220px]]
[[Random forest]]s (RF) classify by constructing an ensemble of [[decision trees]], and outputting the average prediction of the individual trees.<ref>{{cite conference | vauthors = Ho TK |title=Random Decision Forests | conference =Proceedings of the 3rd International Conference on Document Analysis and Recognition, Montreal, QC, 14–16 August 1995. |date=1995 |pages=278–282}}</ref> This is a modification of [[bootstrap aggregating]] (which aggregates a large collection of decision trees) and can be used for [[classification]] or [[regression analysis|regression]].<ref>{{cite book| vauthors = Dietterich T |title=An Experimental Comparison of Three Methodsfor Constructing Ensembles of Decision Trees:Bagging, Boosting, and Randomization |date=2000 |publisher=Kluwer Academic Publishers |pages=139–157}}</ref><ref>{{Cite journal |last=Breiman |first=Leo |date=2001 |title=Radom Forests |journal=Machine Learning |volume=45 |issue=1 |pages=5–32 |doi=10.1023/A:1010933404324|bibcode=2001MachL..45....5B |s2cid=89141 |doi-access=free }}</ref>

As random forests give an internal estimate of generalization error, cross-validation is unnecessary. In addition, they produce proximities, which can be used to impute missing values, and which enable novel data visualizations.<ref name = "Zhang_2012">{{cite book| vauthors = Zhang C, Ma Y |title=Ensemble machine learning: methods and applications |date=2012 |publisher=Springer New York Dordrecht Heidelberg London |location=New York |isbn=978-1-4419-9325-0 |pages=157–175}}</ref>

Computationally, random forests are appealing because they naturally handle both regression and (multiclass) classification, are relatively fast to train and to predict, depend only on one or two tuning parameters, have a built-in estimate of the generalization error, can be used directly for high-dimensional problems, and can easily be implemented in parallel. Statistically, random forests are appealing for additional features, such as measures of variable importance, differential class weighting, missing value imputation, visualization, outlier detection, and unsupervised learning.<ref name = "Zhang_2012" />

=== Clustering ===
[[Cluster analysis|Clustering]] - the partitioning of a data set into disjoint subsets, so that the data in each subset are as close as possible to each other and as distant as possible from data in any other subset, according to some defined [[Metric (mathematics)|distance]] or [[Similarity measure|similarity]] function - is a common technique for statistical data analysis.

Clustering is central to much data-driven bioinformatics research and serves as a powerful computational method whereby means of hierarchical, centroid-based, distribution-based, density-based, and self-organizing maps classification, has long been studied and used in classical machine learning settings. Particularly, clustering helps to analyze unstructured and high-dimensional data in the form of sequences, expressions, texts, images, and so on. Clustering is also used to gain insights into biological processes at the [[genomic]] level, e.g. gene functions, cellular processes, subtypes of cells, [[gene regulation]], and metabolic processes.<ref>{{cite journal | vauthors = Karim MR, Beyan O, Zappa A, Costa IG, Rebholz-Schuhmann D, Cochez M, Decker S | title = Deep learning-based clustering approaches for bioinformatics | journal = Briefings in Bioinformatics | volume = 22 | issue = 1 | pages = 393–415 | date = January 2021 | pmid = 32008043 | pmc = 7820885 | doi = 10.1093/bib/bbz170 }}</ref>

==== Clustering algorithms used in bioinformatics ====
Data clustering algorithms can be hierarchical or partitional. Hierarchical algorithms find successive clusters using previously established clusters, whereas partitional algorithms determine all clusters at once. Hierarchical algorithms can be agglomerative (bottom-up) or divisive (top-down).

Agglomerative algorithms begin with each element as a separate cluster and merge them in successively larger clusters. Divisive algorithms begin with the whole set and proceed to divide it into successively smaller clusters. [[Hierarchical clustering]] is calculated using metrics on [[Euclidean spaces]], the most commonly used is the [[Euclidean distance]] computed by finding the square of the difference between each variable, adding all the squares, and finding the square root of the said sum. An example of a [[hierarchical clustering]] algorithm is [[BIRCH]], which is particularly good on bioinformatics for its nearly linear [[time complexity]] given generally large datasets.<ref name="Variations on the Clustering Algori">{{cite journal |title=Variations on the Clustering Algorithm BIRCH |journal=Big Data Research |date=March 1, 2018 |volume=11 |pages=44–53 |doi=10.1016/j.bdr.2017.09.002| vauthors = Lorbeer B, Kosareva A, Deva B, Softić D, Ruppel P, Küpper A |doi-access=free }}</ref> Partitioning algorithms are based on specifying an initial number of groups, and iteratively reallocating objects among groups to convergence. This algorithm typically determines all clusters at once. Most applications adopt one of two popular heuristic methods: [[k-means]] algorithm or [[k-medoids]]. Other algorithms do not require an initial number of groups, such as [[affinity propagation]]. In a genomic setting this algorithm has been used both to cluster biosynthetic gene clusters in gene cluster families(GCF) and to cluster said GCFs.<ref>{{cite journal | vauthors = Navarro-Muñoz JC, Selem-Mojica N, Mullowney MW, Kautsar SA, Tryon JH, Parkinson EI, De Los Santos EL, Yeong M, Cruz-Morales P, Abubucker S, Roeters A, Lokhorst W, Fernandez-Guerra A, Cappelini LT, Goering AW, Thomson RJ, Metcalf WW, Kelleher NL, Barona-Gomez F, Medema MH | display-authors = 6 | title = A computational framework to explore large-scale biosynthetic diversity | journal = Nature Chemical Biology | volume = 16 | issue = 1 | pages = 60–68 | date = January 2020 | pmid =  31768033| pmc =  6917865| doi = 10.1038/s41589-019-0400-9 }}</ref>

=== Workflow ===

Typically, a workflow for applying machine learning to biological data goes through four steps:<ref name="Larrañaga2006"/> 
* ''Recording,'' including capture and storage. In this step, different information sources may be merged into a single set.
* ''Preprocessing,'' including cleaning and restructuring into a ready-to-analyze form. In this step, uncorrected data are eliminated or corrected, while missing data maybe imputed and relevant variables chosen.
* ''Analysis,'' evaluating data using either supervised or unsupervised algorithms. The algorithm is typically trained on a subset of data, optimizing parameters, and evaluated on a separate test subset.
* ''Visualization and interpretation,'' where knowledge is represented effectively using different methods to assess the significance and importance of the findings.

=== Data errors ===
* Duplicate data is a significant issue in bioinformatics. Publicly available data may be of uncertain quality.<ref name="Shastry2020" >{{cite book | vauthors = Shastry KA, Sanjay HA |title=Statistical Modelling and Machine Learning Principles for Bioinformatics Techniques, Tools, and Applications |publisher=Springer Singapore |isbn=978-981-15-2444-8 |pages=25–39 |chapter-url=https://link-springer-com.pbidi.unam.mx:2443/chapter/10.1007/978-981-15-2445-5_3 |chapter=Machine Learning for Bioinformatics|series=Algorithms for Intelligent Systems |year=2020 |doi=10.1007/978-981-15-2445-5_3 |s2cid=214350490 }}</ref>
* Errors during experimentation.<ref name="Shastry2020"/>
* Erroneous interpretation.<ref name="Shastry2020"/>
* Typing mistakes.<ref name="Shastry2020"/>
* Non-standardized methods (3D structure in PDB from multiple sources, X-ray diffraction, theoretical modeling, nuclear magnetic resonance, etc.) are used in experiments.<ref name="Shastry2020"/>

== Applications ==
In general, a machine learning system can usually be trained to recognize elements of a certain class given sufficient samples.<ref name="Libbrecht2015">{{cite journal|vauthors=Libbrecht MW, Noble WS|date=June 2015|title=Machine learning applications in genetics and genomics|journal=Nature Reviews. Genetics|volume=16|issue=6|pages=321–32|doi=10.1038/nrg3920|pmc=5204302|pmid=25948244}}</ref> For example, machine learning methods can be trained to identify specific visual features such as splice sites.<ref name="Degroeve2002">{{cite journal|vauthors=Degroeve S, De Baets B, Van de Peer Y, Rouzé P|date=2002|title=Feature subset selection for splice site prediction|journal=Bioinformatics|volume=18|issue=Suppl 2 |pages=S75-83|doi=10.1093/bioinformatics/18.suppl_2.s75|pmid=12385987|doi-access=free}}</ref>

[[Support-vector machine|Support vector machines]] have been extensively used in cancer genomic studies.<ref name="Huang2018">{{cite journal|vauthors=Huang S, Cai N, Pacheco PP, Narrandes S, Wang Y, Xu W|date=January 2018|title=Applications of Support Vector Machine (SVM) Learning in Cancer Genomics|journal=Cancer Genomics & Proteomics|volume=15|issue=1|pages=41–51|doi=10.21873/cgp.20063|pmc=5822181|pmid=29275361}}</ref> In addition, [[deep learning]] has been incorporated into bioinformatic algorithms. Deep learning applications have been used for regulatory genomics and cellular imaging.<ref name="Angermueller2016">{{cite journal|vauthors=Angermueller C, Pärnamaa T, Parts L, Stegle O|date=July 2016|title=Deep learning for computational biology|journal=Molecular Systems Biology|volume=12|issue=7|pages=878|doi=10.15252/msb.20156651|pmc=4965871|pmid=27474269}}</ref> Other applications include medical image classification, genomic sequence analysis, as well as protein structure classification and prediction.<ref name="Cao2018">{{cite journal|display-authors=6|vauthors=Cao C, Liu F, Tan H, Song D, Shu W, Li W, Zhou Y, Bo X, Xie Z|date=February 2018|title=Deep Learning and Its Applications in Biomedicine|journal=Genomics, Proteomics & Bioinformatics|volume=16|issue=1|pages=17–32|doi=10.1016/j.gpb.2017.07.003|pmc=6000200|pmid=29522900}}</ref> Deep learning has been applied to regulatory genomics, variant calling and pathogenicity scores.<ref name="Zou2019">{{cite journal|vauthors=Zou J, Huss M, Abid A, Mohammadi P, Torkamani A, Telenti A|date=January 2019|title=A primer on deep learning in genomics|journal=Nature Genetics|volume=51|issue=1|pages=12–18|doi=10.1038/s41588-018-0295-5|pmid=30478442|s2cid=205572042|pmc=11180539}}</ref> [[Natural language processing]] and [[text mining]] have helped to understand phenomena including protein-protein interaction, gene-disease relation as well as predicting biomolecule structures and functions.<ref>{{cite journal|vauthors=Zeng Z, Shi H, Wu Y, Hong Z|date=2015|title=Survey of Natural Language Processing Techniques in Bioinformatics|journal=Computational and Mathematical Methods in Medicine|volume=2015|issue=D1|pages=674296|doi=10.1155/2015/674296|pmc=4615216|pmid=26525745|doi-access=free}}</ref>

=== Precision/personalized medicine ===
[[Natural Language Processing|Natural language processing]] algorithms [[personalized medicine]] for patients who suffer genetic diseases, by combining the extraction of clinical information and genomic data available from the patients. Institutes such as Health-funded Pharmacogenomics Research Network focus on finding breast cancer treatments.<ref>{{cite journal|vauthors=Zeng Z, Shi H, Wu Y, Hong Z|date=2012|title=Survey of Natural Language Processing Techniques in Bioinformatics|journal=Computational and Mathematical Methods in Medicine|volume=2015|issue=D1|pages=674296|doi=10.1016/B978-0-12-385467-4.00006-3|pmc=4615216|pmid=26525745}}</ref>

[[Precision medicine]] considers individual genomic variability, enabled by large-scale biological databases. Machine learning can be applied to perform the matching function between (groups of patients) and specific treatment modalities.<ref>{{cite journal|vauthors=Zeng Z, Shi H, Wu Y, Hong Z|date=2017|title=Survey of Natural Language Processing Techniques in Bioinformatics|journal=Computational and Mathematical Methods in Medicine|volume=2015|issue=D1|pages=674296|doi=10.1155/2015/674296|pmc=4615216|pmid=26525745|doi-access=free}}</ref>

Computational techniques are used to solve other problems, such as efficient primer design for [[Polymerase chain reaction|PCR]], biological-image analysis and back translation of proteins (which is, given the degeneration of the genetic code, a complex combinatorial problem).<ref name="Larrañaga2006" />

=== Genomics ===
[[File:Growth of GenBank.png|thumb|The growth of GenBank, a genomic sequence database provided by the National center for Biotechnology Information (NCBI)]]
While genomic sequence data has historically been sparse due to the technical difficulty of sequencing a piece of DNA, the number of available sequences is growing. On average, the number of [[Base (genetics)|bases]] available in the [[GenBank]] public repository has doubled every 18 months since 1982.<ref>{{Cite web |title=GenBank and WGS Statistics |url=https://www.ncbi.nlm.nih.gov/genbank/statistics/ |access-date=2023-11-25 |website=www.ncbi.nlm.nih.gov}}</ref>  However, while [[raw data]] was becoming increasingly available and accessible, {{As of|2002}}, biological interpretation of this data was occurring at a much slower pace.<ref name=":1">{{cite journal | vauthors = Mathé C, Sagot MF, Schiex T, Rouzé P | title = Current methods of gene prediction, their strengths and weaknesses | journal = Nucleic Acids Research | volume = 30 | issue = 19 | pages = 4103–17 | date = October 2002 | pmid = 12364589 | pmc = 140543 | doi = 10.1093/nar/gkf543 }}</ref> This made for an increasing need for developing [[computational genomics]] tools, including machine learning systems, that can automatically determine the location of protein-encoding genes within a given DNA sequence (i.e. [[gene prediction]]).<ref name=":1" />

Gene prediction is commonly performed through both ''extrinsic searches'' and ''intrinsic searches''.<ref name=":1" /> For the extrinsic search, the input DNA sequence is run through a large database of sequences whose genes have been previously discovered and their locations annotated and identifying the target sequence's genes by determining which strings of bases within the sequence are [[Homology (biology)|homologous]] to known gene sequences. However, not all the genes in a given input sequence can be identified through homology alone, due to limits in the size of the database of known and annotated gene sequences. Therefore, an intrinsic search is needed where a gene prediction program attempts to identify the remaining genes from the DNA sequence alone.<ref name=":1" />

Machine learning has also been used for the problem of [[multiple sequence alignment]] which involves aligning many DNA or amino acid sequences in order to determine regions of similarity that could indicate a shared evolutionary history.<ref name="Larrañaga2006"/> It can also be used to detect and visualize genome rearrangements.<ref name="rearrang">{{cite journal | vauthors = Pratas D, Silva RM, Pinho AJ, Ferreira PJ | title = An alignment-free method to find and visualise rearrangements between pairs of DNA sequences | journal = Scientific Reports | volume = 5 | pages = 10203 | date = May 2015 | pmid = 25984837 | pmc = 4434998 | doi = 10.1038/srep10203 | number = 10203 | bibcode = 2015NatSR...510203P }}</ref>

=== Proteomics ===
[[File:C16orf95 protein secondary structure prediction.png|thumb|A protein's amino acid sequence annotated with the protein secondary structure. Each amino acid is labeled as either an alpha helix, a beta-sheet, or a coil.]]
[[Protein]]s, strings of [[amino acid]]s, gain much of their function from [[protein folding]], where they conform into a three-dimensional structure, including the [[Protein primary structure|primary structure]], the [[Protein secondary structure|secondary structure]] ([[Alpha helix|alpha helices]] and [[beta sheet]]s), the [[Protein tertiary structure|tertiary structure]], and the [[Protein quaternary structure|quaternary structure]].

Protein secondary structure prediction is a main focus of this subfield as tertiary and quaternary structures are determined based on the secondary structure.<ref name=":2" /> Solving the true structure of a protein is expensive and time-intensive, furthering the need for systems that can accurately predict the structure of a protein by analyzing the amino acid sequence directly.<ref name=":2" /><ref name="Larrañaga2006"/> Prior to machine learning, researchers needed to conduct this prediction manually. This trend began in 1951 when Pauling and Corey released their work on predicting the hydrogen bond configurations of a protein from a polypeptide chain.<ref>{{cite journal | vauthors = Pauling L, Corey RB, Branson HR | title = The structure of proteins; two hydrogen-bonded helical configurations of the polypeptide chain | journal = Proceedings of the National Academy of Sciences of the United States of America | volume = 37 | issue = 4 | pages = 205–11 | date = April 1951 | pmid = 14816373 | pmc = 1063337 | doi = 10.1073/pnas.37.4.205 | bibcode = 1951PNAS...37..205P | doi-access = free }}</ref> Automatic feature learning reaches an accuracy of 82-84%.<ref name=":2" /><ref name=":3" /> The current state-of-the-art in secondary structure prediction uses a system called DeepCNF (deep convolutional neural fields) which relies on the machine learning model of [[artificial neural network]]s to achieve an accuracy of approximately 84% when tasked to classify the amino acids of a protein sequence into one of three structural classes (helix, sheet, or coil).<ref name=":3">{{cite journal | vauthors = Wang S, Peng J, Ma J, Xu J | title = Protein Secondary Structure Prediction Using Deep Convolutional Neural Fields | journal = Scientific Reports | volume = 6 | pages = 18962 | date = January 2016 | pmid = 26752681 | pmc = 4707437 | doi = 10.1038/srep18962 | arxiv = 1512.00843 | bibcode = 2016NatSR...618962W }}</ref> {{Update inline|date=June 2021}} The theoretical limit for three-state protein secondary structure is 88–90%.<ref name=":2" />

Machine learning has also been applied to proteomics problems such as [[Side chain|protein side-chain]] prediction, [[Turn (biochemistry)|protein loop]] modeling, and [[protein contact map]] prediction.<ref name="Larrañaga2006"/>

=== Metagenomics ===
[[Metagenomics]] is the study of microbial communities from environmental DNA samples.<ref>{{cite journal | vauthors = Riesenfeld CS, Schloss PD, Handelsman J | title = Metagenomics: genomic analysis of microbial communities | journal = Annual Review of Genetics | volume = 38 | issue = 1 | pages = 525–52 | date = 2004 | pmid = 15568985 | doi = 10.1146/annurev.genet.38.072902.091216 }}</ref> Currently, limitations and challenges predominate in the implementation of machine learning tools due to the amount of data in environmental samples.<ref>{{Cite journal |last1=Soueidan |first1=Hayssam |last2=Nikolski |first2=Macha |date=2017-01-01 |title=Machine learning for metagenomics: methods and tools |url=http://access.portico.org/stable?au=phz1pprtrsb |journal=Metagenomics |volume=1 |issue=1 |doi=10.1515/metgen-2016-0001 |s2cid=17418188 |issn=2449-7657|doi-access=free |arxiv=1510.06621 }}</ref> Supercomputers and web servers have made access to these tools easier.<ref name = "Lin_2021">{{cite journal | vauthors = Lin Y, Wang G, Yu J, Sung JJ | title = Artificial intelligence and metagenomics in intestinal diseases | journal = Journal of Gastroenterology and Hepatology | volume = 36 | issue = 4 | pages = 841–847 | date = April 2021 | pmid = 33880764 | doi = 10.1111/jgh.15501 | s2cid = 233312307 | doi-access = free }}</ref> The high dimensionality of microbiome datasets is a major challenge in studying the microbiome; this significantly limits the power of current approaches for identifying true differences and increases the chance of false discoveries.<ref name = "Dang_2020">{{cite bioRxiv | vauthors = Dang T, Kishino H | title = Detecting significant components of microbiomes by random forest with forward variable selection and phylogenetics | date = January 2020 | biorxiv = 10.1101/2020.10.29.361360 }}</ref>{{better source needed|date=November 2023}}

Despite their importance, machine learning tools related to metagenomics have focused on the study of gut microbiota and the relationship with digestive diseases, such as [[inflammatory bowel disease]] (IBD), ''[[Clostridioides difficile]]'' infection (CDI), [[colorectal cancer]] and [[diabetes]], seeking better diagnosis and treatments.<ref name = "Lin_2021" /> Many algorithms were developed to classify microbial communities according to the health condition of the host, regardless of the type of sequence data, e.g. [[16S rRNA]] or [[whole-genome sequencing]] (WGS), using methods such as least absolute shrinkage and selection operator classifier, [[random forest]], supervised classification model, and gradient boosted tree model. [[Neural networks]], such as [[recurrent neural network]]s (RNN), [[convolutional neural network]]s (CNN), and [[Hopfield neural network]]s have been added.<ref name = "Lin_2021" /> For example, in 2018, Fioravanti et al. developed an algorithm called Ph-CNN to classify data samples from healthy patients and patients with IBD symptoms (to distinguish healthy and sick patients) by using phylogenetic trees and convolutional neural networks.<ref name = "Fioravanti2018">{{cite journal | vauthors = Fioravanti D, Giarratano Y, Maggio V, Agostinelli C, Chierici M, Jurman G, Furlanello C | title = Phylogenetic convolutional neural networks in metagenomics | journal = BMC Bioinformatics | volume = 19 | issue = Suppl 2 | pages = 49 | date = March 2018 | pmid = 29536822 | pmc = 5850953 | doi = 10.1186/s12859-018-2033-5 | doi-access = free }}</ref>

In addition, [[random forest]] (RF) methods and implemented importance measures help in the identification of microbiome species that can be used to distinguish diseased and non-diseased samples. However, the performance of a decision tree and the diversity of decision trees in the ensemble significantly influence the performance of RF algorithms. The generalization error for RF measures how accurate the individual classifiers are and their interdependence. Therefore, the high dimensionality problems of microbiome datasets pose challenges. Effective approaches require many possible variable combinations, which exponentially increases the computational burden as the number of features increases.<ref name = "Dang_2020" />

For microbiome analysis in 2020 Dang & Kishino<ref name="Dang_2020" /> developed a novel analysis pipeline. The core of the pipeline is an RF classifier coupled with forwarding [[variable selection]] (RF-FVS), which selects a minimum-size core set of microbial species or functional signatures that maximize the predictive classifier performance. The framework combines:

* identifying a few significant features by a massively parallel forward [[variable selection]] procedure
* mapping the selected species on a [[phylogenetic tree]], and
* predicting functional profiles by functional gene enrichment analysis from metagenomic [[16S rRNA]] data.

They demonstrated performance by analyzing two published datasets from large-scale case-control studies:

* 16S rRNA gene amplicon data for ''C. difficile'' infection (CDI) and 
* shotgun metagenomics data for human colorectal cancer (CRC).

The proposed approach improved the accuracy from 81% to 99.01% for CDI and from 75.14% to 90.17% for CRC.

The use of machine learning in environmental samples has been less explored, maybe because of data complexity, especially from WGS. Some works show that it is possible to apply these tools in environmental samples. In 2021 Dhungel et al.,<ref>{{cite journal | vauthors = Dhungel E, Mreyoud Y, Gwak HJ, Rajeh A, Rho M, Ahn TH | title = MegaR: an interactive R package for rapid sample classification and phenotype prediction using metagenome profiles and machine learning | journal = BMC Bioinformatics | volume = 22 | issue = 1 | pages = 25 | date = January 2021 | pmid = 33461494 | pmc = 7814621 | doi = 10.1186/s12859-020-03933-4 | doi-access = free }}</ref> designed an R package called MegaR. This package allows working with 16S rRNA and whole metagenomic sequences to make taxonomic profiles and classification models by machine learning models. MegaR includes a comfortable visualization environment to improve the user experience. Machine learning in environmental metagenomics can help to answer questions related to the interactions between microbial communities and ecosystems, e.g. the work of Xun et al., in 2021<ref>{{cite journal | vauthors = Xun W, Liu Y, Li W, Ren Y, Xiong W, Xu Z, Zhang N, Miao Y, Shen Q, Zhang R | display-authors = 6 | title = Specialized metabolic functions of keystone taxa sustain soil microbiome stability | journal = Microbiome | volume = 9 | issue = 1 | pages = 35 | date = January 2021 | pmid = 33517892 | pmc = 7849160 | doi = 10.1186/s40168-020-00985-9 | doi-access = free }}</ref> where the use of different machine learning methods offered insights on the relationship among the soil, microbiome biodiversity, and ecosystem stability.

=== Microarrays ===
[[Microarray]]s, a type of [[lab-on-a-chip]], are used for automatically collecting data about large amounts of biological material. Machine learning can aid in analysis, and has been applied to expression pattern identification, classification, and genetic network induction.<ref name="Larrañaga2006"/>
[[File:DNA-microarray analysis.jpg|thumb|A DNA-microarray analysis of Burkitt's lymphoma and diffuse large B-cell lymphoma (DLBCL), which differences in gene expression patterns]]
This technology is especially useful for monitoring gene expression, aiding in diagnosing cancer by examining which genes are expressed.<ref name=":4">{{cite journal | vauthors = Pirooznia M, Yang JY, Yang MQ, Deng Y | title = A comparative study of different machine learning methods on microarray gene expression data | journal = BMC Genomics | volume = 9 Suppl 1 | issue = 1 | pages = S13 | date = 2008 | pmid = 18366602 | pmc = 2386055 | doi = 10.1186/1471-2164-9-S1-S13 | doi-access = free }}</ref> One of the main tasks is identifying which genes are expressed based on the collected data.<ref name="Larrañaga2006"/> In addition, due to the huge number of genes on which data is collected by the microarray, winnowing the large amount of irrelevant data to the task of expressed gene identification is challenging. Machine learning presents a potential solution as various classification methods can be used to perform this identification. The most commonly used methods are [[radial basis function network]]s, [[deep learning]], [[Naive Bayes classifier|Bayesian classification]], [[decision tree]]s, and [[random forest]].<ref name=":4" />

=== Systems biology ===
Systems biology focuses on the study of emergent behaviors from complex interactions of simple biological components in a system. Such components can include DNA, RNA, proteins, and metabolites.<ref>{{cite web|url=http://journal.frontiersin.org/researchtopic/2362/machine-learning-in-molecular-systems-biology|title=Machine Learning in Molecular Systems Biology|website=Frontiers|language=en|access-date=2017-06-09}}</ref>

Machine learning has been used to aid in modeling these interactions in domains such as genetic networks, signal transduction networks, and metabolic pathways.<ref name="Larrañaga2006"/> [[Graphical model|Probabilistic graphical models]], a machine learning technique for determining the relationship between different variables, are one of the most commonly used methods for modeling genetic networks.<ref name="Larrañaga2006"/> In addition, machine learning has been applied to systems biology problems such as identifying [[DNA binding site|transcription factor binding sites]] using [[Markov chain|Markov chain optimization]].<ref name="Larrañaga2006"/> [[Genetic algorithm]]s, machine learning techniques which are based on the natural process of evolution, have been used to model genetic networks and regulatory structures.<ref name="Larrañaga2006"/>

Other systems biology applications of machine learning include the task of enzyme function prediction, high throughput microarray data analysis, analysis of genome-wide association studies to better understand markers of disease, protein function prediction.<ref>{{cite journal | vauthors = d'Alché-Buc F, Wehenkel L | title = Machine learning in systems biology | journal = BMC Proceedings | volume = 2 Suppl 4 | issue = 4 | pages = S1 | date = December 2008 | pmid = 19091048 | pmc = 2654969 | doi = 10.1186/1753-6561-2-S4-S1 | doi-access = free }}</ref>

=== Evolution ===
This domain, particularly [[phylogenetic tree]] reconstruction, uses the features of machine learning techniques. Phylogenetic trees are schematic representations of the evolution of organisms. Initially, they were constructed using features such as morphological and metabolic features. Later, due to the availability of genome sequences, the construction of the phylogenetic tree algorithm used the concept based on genome comparison. With the help of optimization techniques, a comparison was done by means of multiple sequence alignment.<ref>{{cite book | vauthors = Bhattacharya M | veditors = Srinivasa MG, Siddesh GM, MAnisekhar SR |title=Statistical Modelling and Machine Learning Principles for Bioinformatics Techniques, Tools, and Applications |publisher=Springer Singapore |date=2020 |pages=164–188 |chapter= Unsupervised Techniques in Genomics |isbn=978-981-15-2445-5}}</ref>

=== Stroke diagnosis ===
Machine learning methods for the analysis of [[neuroimaging]] data are used to help diagnose [[stroke]]. Historically multiple approaches to this problem involved neural networks.<ref name = "stroke1">{{cite journal | vauthors = Topol EJ | title = High-performance medicine: the convergence of human and artificial intelligence | journal = Nature Medicine | volume = 25 | issue = 1 | pages = 44–56 | date = January 2019 | pmid = 30617339 | doi = 10.1038/s41591-018-0300-7 | s2cid = 57574615 | hdl = 10654/45728 | hdl-access = free }}</ref><ref name="stroke2">{{cite journal | vauthors = Jiang F, Jiang Y, Zhi H, Dong Y, Li H, Ma S, Wang Y, Dong Q, Shen H, Wang Y | display-authors = 6 | title = Artificial intelligence in healthcare: past, present and future | journal = Stroke and Vascular Neurology | volume = 2 | issue = 4 | pages = 230–243 | date = December 2017 | pmid = 29507784 | pmc = 5829945 | doi = 10.1136/svn-2017-000101 }}</ref>

Multiple approaches to detect strokes used machine learning. As proposed by Mirtskhulava,<ref>{{cite book | vauthors = Mirtskhulava L, Wong J, Al-Majeed S, Pearce G  |title=2015 17th UKSim-AMSS International Conference on Modelling and Simulation (UKSim) |chapter=Artificial Neural Network Model in Stroke Diagnosis |date=March 2015 |pages=50–53 |doi=10.1109/UKSim.2015.33|isbn=978-1-4799-8713-9 |s2cid=6391733 |chapter-url=http://eprints.glos.ac.uk/5581/1/5581%20Al-Majeed%20%282015%29%20Artificial%20Neural%20Network%20Model%20in%20Strole%20Diagnosis.pdf }}</ref> feed-forward networks were tested to detect strokes using neural imaging. As proposed by Titano<ref>{{cite journal | vauthors = Titano JJ, Badgeley M, Schefflein J, Pain M, Su A, Cai M, Swinburne N, Zech J, Kim J, Bederson J, Mocco J, Drayer B, Lehar J, Cho S, Costa A, Oermann EK | display-authors = 6 | title = Automated deep-neural-network surveillance of cranial images for acute neurologic events | journal = Nature Medicine | volume = 24 | issue = 9 | pages = 1337–1341 | date = September 2018 | pmid = 30104767 | doi = 10.1038/s41591-018-0147-y | s2cid = 51976344 }}</ref> 3D-CNN techniques were tested in supervised classification to screen head CT images for acute neurologic events. Three-dimensional [[Convolutional neural network|CNN]] and [[Support-vector machine|SVM]] methods are often used.<ref name="stroke2"/>

=== Text mining ===
The increase in biological publications increased the difficulty in searching and compiling relevant available information on a given topic. This task is known as [[knowledge extraction]]. It is necessary for biological data collection which can then in turn be fed into machine learning algorithms to generate new biological knowledge.<ref name="Larrañaga2006"/><ref name=":5">{{cite journal | vauthors = Krallinger M, Erhardt RA, Valencia A | title = Text-mining approaches in molecular biology and biomedicine | journal = Drug Discovery Today | volume = 10 | issue = 6 | pages = 439–45 | date = March 2005 | pmid = 15808823 | doi = 10.1016/S1359-6446(05)03376-3 }}</ref> Machine learning can be used for this knowledge extraction task using techniques such as [[natural language processing]] to extract the useful information from human-generated reports in a database. [[Text Nailing]], an alternative approach to machine learning, capable of extracting features from clinical narrative notes was introduced in 2017.

This technique has been applied to the search for novel drug targets, as this task requires the examination of information stored in biological databases and journals.<ref name=":5" /> Annotations of proteins in protein databases often do not reflect the complete known set of knowledge of each protein, so additional information must be extracted from biomedical literature. Machine learning has been applied to the automatic annotation of gene and protein function, determination of the [[Protein targeting|protein subcellular localization]], [[Gene expression|DNA-expression array]] analysis, large-scale [[List of protein interactions|protein interaction]] analysis, and molecule interaction analysis.<ref name=":5" />

Another application of text mining is the detection and visualization of distinct DNA regions given sufficient reference data.<ref name="sing">{{cite book| vauthors = Pratas D, Hosseini M, Silva R, Pinho A, Ferreira P |title=Pattern Recognition and Image Analysis|chapter=Visualization of Distinct DNA Regions of the Modern Human Relatively to a Neanderthal Genome|volume=10255|pages=235–242|date= June 20–23, 2017|doi=10.1007/978-3-319-58838-4_26|series=Lecture Notes in Computer Science|isbn=978-3-319-58837-7}}</ref>

=== Clustering and abundance profiling of biosynthetic gene clusters===

Microbial communities are complex assembles of diverse microorganisms,<ref>{{cite journal | vauthors = Bardgett RD, Caruso T | title = Soil microbial community responses to climate extremes: resistance, resilience and transitions to alternative states | journal = Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences | volume = 375 | issue = 1794 | pages = 20190112 | date = March 2020 | pmid = 31983338 | pmc = 7017770 | doi = 10.1098/rstb.2019.0112 }}</ref> where symbiont partners constantly produce diverse metabolites derived from the primary and secondary (specialized) metabolism, from which metabolism plays an important role in microbial interaction.<ref>{{cite journal | vauthors = Deveau A, Bonito G, Uehling J, Paoletti M, Becker M, Bindschedler S, Hacquard S, Hervé V, Labbé J, Lastovetsky OA, Mieszkin S, Millet LJ, Vajna B, Junier P, Bonfante P, Krom BP, Olsson S, van Elsas JD, Wick LY | display-authors = 6 | title = Bacterial-fungal interactions: ecology, mechanisms and challenges | journal = FEMS Microbiology Reviews | volume = 42 | issue = 3 | pages = 335–352 | date = May 2018 | pmid = 29471481 | doi = 10.1093/femsre/fuy008 | doi-access = free | hdl = 21.11116/0000-0002-C1E7-F | hdl-access = free }}</ref> Metagenomic and metatranscriptomic data are an important source for deciphering communications signals.

Molecular mechanisms produce specialized metabolites in various ways. [[Metabolic gene cluster|Biosynthetic Gene Clusters]] (BGCs) attract attention, since several metabolites are clinically valuable, anti-microbial, anti-fungal, anti-parasitic, anti-tumor and immunosuppressive agents produced by the modular action of multi-enzymatic, multi-domains gene clusters, such as [[Nonribosomal peptide]] synthetases (NRPSs) and [[polyketide synthase]]s (PKSs).<ref>{{cite journal | vauthors = Ansari MZ, Yadav G, Gokhale RS, Mohanty D | title = NRPS-PKS: a knowledge-based resource for analysis of NRPS/PKS megasynthases | journal = Nucleic Acids Research | volume = 32 | issue = Web Server issue | pages = W405-13 | date = July 2004 | pmid = 15215420 | pmc = 441497 | doi = 10.1093/nar/gkh359 }}</ref> Diverse studies<ref name="A computational framework to explor">{{cite journal | vauthors = Navarro-Muñoz JC, Selem-Mojica N, Mullowney MW, Kautsar SA, Tryon JH, Parkinson EI, De Los Santos EL, Yeong M, Cruz-Morales P, Abubucker S, Roeters A, Lokhorst W, Fernandez-Guerra A, Cappelini LT, Goering AW, Thomson RJ, Metcalf WW, Kelleher NL, Barona-Gomez F, Medema MH | display-authors = 6 | title = A computational framework to explore large-scale biosynthetic diversity | journal = Nature Chemical Biology | volume = 16 | issue = 1 | pages = 60–68 | date = January 2020 | pmid = 31768033 | pmc = 6917865 | doi = 10.1038/s41589-019-0400-9 }}</ref><ref name="A roadmap for natural product disco">{{cite journal | vauthors = Doroghazi JR, Albright JC, Goering AW, Ju KS, Haines RR, Tchalukov KA, Labeda DP, Kelleher NL, Metcalf WW | display-authors = 6 | title = A roadmap for natural product discovery based on large-scale genomics and metabolomics | journal = Nature Chemical Biology | volume = 10 | issue = 11 | pages = 963–8 | date = November 2014 | pmid = 25262415 | pmc = 4201863 | doi = 10.1038/nchembio.1659 }}</ref><ref name="Insights into secondary metabolism">{{cite journal | vauthors = Cimermancic P, Medema MH, Claesen J, Kurita K, Wieland Brown LC, Mavrommatis K, Pati A, Godfrey PA, Koehrsen M, Clardy J, Birren BW, Takano E, Sali A, Linington RG, Fischbach MA | display-authors = 6 | title = Insights into secondary metabolism from a global analysis of prokaryotic biosynthetic gene clusters | journal = Cell | volume = 158 | issue = 2 | pages = 412–421 | date = July 2014 | pmid = 25036635 | pmc = 4123684 | doi = 10.1016/j.cell.2014.06.034 }}</ref><ref name="Metabologenomics: Correlation of Mi">{{cite journal | vauthors = Goering AW, McClure RA, Doroghazi JR, Albright JC, Haverland NA, Zhang Y, Ju KS, Thomson RJ, Metcalf WW, Kelleher NL | display-authors = 6 | title = Metabologenomics: Correlation of Microbial Gene Clusters with Metabolites Drives Discovery of a Nonribosomal Peptide with an Unusual Amino Acid Monomer | journal = ACS Central Science | volume = 2 | issue = 2 | pages = 99–108 | date = February 2016 | pmid = 27163034 | pmc = 4827660 | doi = 10.1021/acscentsci.5b00331 }}</ref><ref>{{cite journal | vauthors = Amiri Moghaddam J, Crüsemann M, Alanjary M, Harms H, Dávila-Céspedes A, Blom J, Poehlein A, Ziemert N, König GM, Schäberle TF | display-authors = 6 | title = Analysis of the Genome and Metabolome of Marine Myxobacteria Reveals High Potential for Biosynthesis of Novel Specialized Metabolites | journal = Scientific Reports | volume = 8 | issue = 1 | pages = 16600 | date = November 2018 | pmid = 30413766 | pmc = 6226438 | doi = 10.1038/s41598-018-34954-y | bibcode = 2018NatSR...816600A }}</ref><ref>{{cite journal | vauthors = Duncan KR, Crüsemann M, Lechner A, Sarkar A, Li J, Ziemert N, Wang M, Bandeira N, Moore BS, Dorrestein PC, Jensen PR | display-authors = 6 | title = Molecular networking and pattern-based genome mining improves discovery of biosynthetic gene clusters and their products from Salinispora species | journal = Chemistry & Biology | volume = 22 | issue = 4 | pages = 460–471 | date = April 2015 | pmid = 25865308 | pmc = 4409930 | doi = 10.1016/j.chembiol.2015.03.010 }}</ref><ref>{{cite journal | vauthors = Nielsen JC, Grijseels S, Prigent S, Ji B, Dainat J, Nielsen KF, Frisvad JC, Workman M, Nielsen J | display-authors = 6 | title = Global analysis of biosynthetic gene clusters reveals vast potential of secondary metabolite production in Penicillium species | journal = Nature Microbiology | volume = 2 | issue = 6 | pages = 17044 | date = April 2017 | pmid = 28368369 | doi = 10.1038/nmicrobiol.2017.44 | s2cid = 22699928 }}</ref><ref>{{cite journal | vauthors = McClure RA, Goering AW, Ju KS, Baccile JA, Schroeder FC, Metcalf WW, Thomson RJ, Kelleher NL | display-authors = 6 | title = Elucidating the Rimosamide-Detoxin Natural Product Families and Their Biosynthesis Using Metabolite/Gene Cluster Correlations | journal = ACS Chemical Biology | volume = 11 | issue = 12 | pages = 3452–3460 | date = December 2016 | pmid = 27809474 | pmc = 5295535 | doi = 10.1021/acschembio.6b00779 }}</ref> show that grouping BGCs that share homologous core genes into gene cluster families (GCFs) can yield useful insights into the chemical diversity of the analyzed strains, and can support linking BGCs to their secondary metabolites.<ref name="A roadmap for natural product disco"/><ref name="Metabologenomics: Correlation of Mi"/> GCFs have been used as functional markers in human health studies<ref>{{cite journal | vauthors = Cao L, Shcherbin E, Mohimani H | title = A Metabolome- and Metagenome-Wide Association Network Reveals Microbial Natural Products and Microbial Biotransformation Products from the Human Microbiota | journal = mSystems | volume = 4 | issue = 4 | pages = e00387–19, /msystems/4/4/msys.00387–19.atom | date = August 2019 | pmid = 31455639 | pmc = 6712304 | doi = 10.1128/mSystems.00387-19 }}</ref><ref>{{cite journal | vauthors = Olm MR, Bhattacharya N, Crits-Christoph A, Firek BA, Baker R, Song YS, Morowitz MJ, Banfield JF | display-authors = 6 | title = Necrotizing enterocolitis is preceded by increased gut bacterial replication, ''Klebsiella'', and fimbriae-encoding bacteria | journal = Science Advances | volume = 5 | issue = 12 | pages = eaax5727 | date = December 2019 | pmid = 31844663 | pmc = 6905865 | doi = 10.1126/sciadv.aax5727 | bibcode = 2019SciA....5.5727O }}</ref> and to study the ability of soil to suppress fungal pathogens.<ref>{{cite journal | vauthors = Carrión VJ, Perez-Jaramillo J, Cordovez V, Tracanna V, de Hollander M, Ruiz-Buck D, Mendes LW, van Ijcken WF, Gomez-Exposito R, Elsayed SS, Mohanraju P, Arifah A, van der Oost J, Paulson JN, Mendes R, van Wezel GP, Medema MH, Raaijmakers JM | display-authors = 6 | title = Pathogen-induced activation of disease-suppressive functions in the endophytic root microbiome | journal = Science | volume = 366 | issue = 6465 | pages = 606–612 | date = November 2019 | pmid = 31672892 | doi = 10.1126/science.aaw9285 | s2cid = 207814746 | bibcode = 2019Sci...366..606C | doi-access = free | hdl = 1887/3188901 | hdl-access = free }}</ref> Given their direct relationship to catalytic enzymes, and compounds produced from their encoded pathways, BGCs/GCFs can serve as a proxy to explore the chemical space of microbial secondary metabolism. Cataloging GCFs in sequenced microbial genomes yields an overview of the existing chemical diversity and offers insights into future priorities.<ref name="A computational framework to explor"/><ref name="Insights into secondary metabolism"/> Tools such as BiG-SLiCE and BIG-MAP<ref name="BiG-MAP: an automated pipeline to p">{{Cite journal |last1=Pascal Andreu |first1=Victória |last2=Augustijn |first2=Hannah E. |last3=van den Berg |first3=Koen |last4=van der Hooft |first4=Justin J. J. |last5=Fischbach |first5=Michael A. |last6=Medema |first6=Marnix H. |date=2021-10-26 |editor-last=Shank |editor-first=Elizabeth Anne |title=BiG-MAP: an Automated Pipeline To Profile Metabolic Gene Cluster Abundance and Expression in Microbiomes |journal=mSystems |language=en |volume=6 |issue=5 |pages=e0093721 |doi=10.1128/mSystems.00937-21 |issn=2379-5077 |pmc=8547482 |pmid=34581602}}</ref> have emerged with the sole purpose of unveiling the importance of BGCs in natural environments.

=== Decodification of RiPPs chemical structures ===

The increase of experimentally characterized [[ribosomally synthesized and post-translationally modified peptides]] (RiPPs), together with the availability of information on their sequence and chemical structure, selected from databases such as BAGEL, BACTIBASE, MIBIG, and THIOBASE, provide the opportunity to develop machine learning tools to decode the chemical structure and classify them.

In 2017, researchers at the National Institute of Immunology of New Delhi, India, developed RiPPMiner<ref>{{cite journal | vauthors = Agrawal P, Khater S, Gupta M, Sain N, Mohanty D | title = RiPPMiner: a bioinformatics resource for deciphering chemical structures of RiPPs based on prediction of cleavage and cross-links | journal = Nucleic Acids Research | volume = 45 | issue = W1 | pages = W80–W88 | date = July 2017 | pmid = 28499008 | pmc = 5570163 | doi = 10.1093/nar/gkx408 }}</ref> software, a bioinformatics resource for decoding RiPP chemical structures by genome mining. The RiPPMiner web server consists of a query interface and the RiPPDB database. RiPPMiner defines 12 subclasses of RiPPs, predicting the cleavage site of the leader peptide and the final cross-link of the RiPP chemical structure.

=== Mass spectral similarity scoring ===

Many tandem mass spectrometry '''('''MS/MS''')''' based metabolomics studies, such as library matching and molecular networking, use spectral similarity as a proxy for structural similarity. Spec2vec<ref>{{cite journal | vauthors = Huber F, Ridder L, Verhoeven S, Spaaks JH, Diblen F, Rogers S, van der Hooft JJ | title = Spec2Vec: Improved mass spectral similarity scoring through learning of structural relationships | journal = PLOS Computational Biology | volume = 17 | issue = 2 | pages = e1008724 | date = February 2021 | pmid = 33591968 | pmc = 7909622 | doi = 10.1371/journal.pcbi.1008724 | bibcode = 2021PLSCB..17E8724H | doi-access = free }}</ref> algorithm provides a new way of spectral similarity score, based on [[Word2vec|Word2Vec]]. Spec2Vec learns fragmental relationships within a large set of spectral data, in order to assess spectral similarities between molecules and to classify unknown molecules through these comparisons.

For systemic annotation, some metabolomics studies rely on fitting measured fragmentation mass spectra to library spectra or contrasting spectra via network analysis. Scoring functions are used to determine the similarity between pairs of fragment spectra as part of these processes. So far, no research has suggested scores that are significantly different from the commonly utilized [[Cosine similarity|cosine-based similarity]].<ref>{{cite journal | vauthors = Huber F, Ridder L, Verhoeven S, Spaaks JH, Diblen F, Rogers S, van der Hooft JJ | title = Spec2Vec: Improved mass spectral similarity scoring through learning of structural relationships | journal = PLOS Computational Biology | volume = 17 | issue = 2 | pages = e1008724 | date = February 2021 | pmid = 33591968 | doi = 10.1371/journal.pcbi.1008724 | pmc = 7909622 | bibcode = 2021PLSCB..17E8724H | doi-access = free }}</ref>

== Databases ==
An important part of bioinformatics is the management of big datasets, known as databases of reference. Databases exist for each type of biological data, for example for biosynthetic gene clusters and metagenomes.

=== General databases by bioinformatics ===

==== National Center for Biotechnology Information ====
The National Center for Biotechnology Information (NCBI)<ref>{{Cite web| author1 = National Center for Biotechnology Information | author2 = U.S. National Library of Medicine |title= National Center for Biotechnology Information|url=https://www.ncbi.nlm.nih.gov/|access-date=2021-07-30|website=ncbi.nlm.nih.gov|language=en}}</ref> provides a large suite of online resources for biological information and data, including the [[GenBank]] nucleic acid sequence database and the [[PubMed]] database of citations and abstracts for published life science journals. Augmenting many of the Web applications are custom implementations of the BLAST program optimized to search specialized data sets. Resources include PubMed Data Management, RefSeq Functional Elements, genome data download, variation services API, Magic-BLAST, QuickBLASTp, and Identical Protein Groups. All of these resources can be accessed through NCBI.<ref>{{cite journal | vauthors = Agarwala R, Barrett T, Beck J, Benson DA, Bollin C, Bolton E, etal | collaboration = NCBI Resource Coordinators | title = Database resources of the National Center for Biotechnology Information | journal = Nucleic Acids Research | volume = 46 | issue = D1 | pages = D8–D13 | date = January 2018 | pmid = 29140470 | pmc = 5753372 | doi = 10.1093/nar/gkx1095 }}</ref>

=== Bioinformatics analysis for biosynthetic gene clusters  ===

==== antiSMASH ====
antiSMASH allows the rapid genome-wide identification, annotation and analysis of secondary metabolite biosynthesis gene clusters in bacterial and fungal genomes. It integrates and cross-links with a large number of in silico [[secondary metabolite]] analysis tools.<ref>{{cite web |title=antiSMASH database |url=https://antismash-db.secondarymetabolites.org/ |website=antismash-db.secondarymetabolites.org}}</ref>

==== gutSMASH ====
gutSMASH is a tool that systematically evaluates bacterial metabolic potential by predicting both known and novel [[anaerobic organism|anaerobic]] metabolic gene clusters (MGCs) from the gut [[microbiome]].

==== MIBiG ====
MIBiG,<ref>{{Cite web|title=MIBiG: Minimum Information about a Biosynthetic Gene cluster|url=https://mibig.secondarymetabolites.org/|access-date=2021-07-30|website=mibig.secondarymetabolites.org|language=en}}</ref> the minimum information about a biosynthetic gene cluster specification, provides a standard for annotations and [[metadata]] on biosynthetic gene clusters and their molecular products. MIBiG is a Genomic Standards Consortium project that builds on the minimum information about any sequence (MIxS) framework.<ref>[https://mibig.secondarymetabolites.org/ MiBiG ]</ref>

MIBiG facilitates the standardized deposition and retrieval of biosynthetic gene cluster data as well as the development of comprehensive comparative analysis tools. It empowers next-generation research on the biosynthesis, chemistry and ecology of broad classes of societally relevant bioactive [[secondary metabolites]], guided by robust experimental evidence and rich metadata components.<ref>{{cite journal | vauthors = Kautsar SA, Blin K, Shaw S, Navarro-Muñoz JC, Terlouw BR, van der Hooft JJ, van Santen JA, Tracanna V, Suarez Duran HG, Pascal Andreu V, Selem-Mojica N, Alanjary M, Robinson SL, Lund G, Epstein SC, Sisto AC, Charkoudian LK, Collemare J, Linington RG, Weber T, Medema MH | display-authors = 6 | title = MIBiG 2.0: a repository for biosynthetic gene clusters of known function | journal = Nucleic Acids Research | volume = 48 | issue = D1 | pages = D454–D458 | date = January 2020 | pmid = 31612915 | pmc = 7145714 | doi = 10.1093/nar/gkz882 }}</ref>

==== SILVA ====
SILVA<ref>{{Cite web|title=Silva|url=https://www.arb-silva.de/|access-date=2021-07-30|website=arb-silva.de}}</ref> is an interdisciplinary project among biologists and computers scientists assembling a complete database of RNA ribosomal (rRNA) sequences of genes, both small ([[16S ribosomal RNA|16S]], [[18S ribosomal RNA|18S]], SSU) and large ([[23S]], [[28S]], LSU) subunits, which belong to the bacteria, archaea and eukarya domains. These data are freely available for academic and commercial use.<ref>{{cite journal | vauthors = Quast C, Pruesse E, Yilmaz P, Gerken J, Schweer T, Yarza P, Peplies J, Glöckner FO | display-authors = 6 | title = The SILVA ribosomal RNA gene database project: improved data processing and web-based tools | journal = Nucleic Acids Research | volume = 41 | issue = Database issue | pages = D590-6 | date = January 2013 | pmid = 23193283 | pmc = 3531112 | doi = 10.1093/nar/gks1219 }}</ref>

==== Greengenes ====
Greengenes<ref>{{Cite web|title=greengenes.secondgenome.com|url=https://greengenes.secondgenome.com/|access-date=2021-07-30|website=greengenes.secondgenome.com}}</ref> is a full-length [[16S rRNA]] gene database that provides chimera screening, standard alignment and a curated taxonomy based on de novo tree inference.<ref>{{cite journal | vauthors = DeSantis TZ, Hugenholtz P, Larsen N, Rojas M, Brodie EL, Keller K, Huber T, Dalevi D, Hu P, Andersen GL | display-authors = 6 | title = Greengenes, a chimera-checked 16S rRNA gene database and workbench compatible with ARB | journal = Applied and Environmental Microbiology | volume = 72 | issue = 7 | pages = 5069–72 | date = July 2006 | pmid = 16820507 | pmc = 1489311 | doi = 10.1128/AEM.03006-05 | bibcode = 2006ApEnM..72.5069D | doi-access = free }}</ref><ref>{{cite journal | vauthors = McDonald D, Price MN, Goodrich J, Nawrocki EP, DeSantis TZ, Probst A, Andersen GL, Knight R, Hugenholtz P | display-authors = 6 | title = An improved Greengenes taxonomy with explicit ranks for ecological and evolutionary analyses of bacteria and archaea | journal = The ISME Journal | volume = 6 | issue = 3 | pages = 610–8 | date = March 2012 | pmid = 22134646 | pmc = 3280142 | doi = 10.1038/ismej.2011.139 | bibcode = 2012ISMEJ...6..610M }}</ref> Overview: 
* 1,012,863 RNA sequences from 92,684 organisms contributed to RNAcentral.
* The shortest sequence has 1,253 nucleotides, the longest 2,368.
* The average length is 1,402 nucleotides.
* Database version: 13.5.

==== Open Tree of Life Taxonomy ====
Open Tree of Life Taxonomy (OTT)<ref>{{Cite web|title=opentree|url=https://tree.opentreeoflife.org/|access-date=2021-07-30|website=tree.opentreeoflife.org}}</ref> aims to build a complete, dynamic, and digitally available Tree of Life by synthesizing published phylogenetic trees along with taxonomic data. Phylogenetic trees have been classified, aligned, and merged. Taxonomies have been used to fill in sparse regions and gaps left by phylogenies. OTT is a base that has been little used for sequencing analyzes of the 16S region, however, it has a greater number of sequences classified taxonomically down to the genus level compared to SILVA and Greengenes. However, in terms of classification at the edge level, it contains a lesser amount of information<ref>{{cite journal | vauthors = Hinchliff CE, Smith SA, Allman JF, Burleigh JG, Chaudhary R, Coghill LM, Crandall KA, Deng J, Drew BT, Gazis R, Gude K, Hibbett DS, Katz LA, Laughinghouse HD, McTavish EJ, Midford PE, Owen CL, Ree RH, Rees JA, Soltis DE, Williams T, Cranston KA | display-authors = 6 | title = Synthesis of phylogeny and taxonomy into a comprehensive tree of life | journal = Proceedings of the National Academy of Sciences of the United States of America | volume = 112 | issue = 41 | pages = 12764–9 | date = October 2015 | pmid = 26385966 | pmc = 4611642 | doi = 10.1073/pnas.1423041112 | bibcode = 2015PNAS..11212764H | doi-access = free }}</ref>

==== Ribosomal Database Project ====
Ribosomal Database Project (RDP)<ref>{{Cite web|title=RDP Release 11 -- Sequence Analysis Tools|url=http://rdp.cme.msu.edu/|access-date=2021-07-30|website=rdp.cme.msu.edu|archive-date=August 19, 2020|archive-url=https://web.archive.org/web/20200819061707/http://rdp.cme.msu.edu/|url-status=dead}}</ref> is a database that provides RNA ribosomal (rRNA) sequences of small subunits of domain bacterial and archaeal ([[16S ribosomal RNA|16S]]); and fungal rRNA sequences of large subunits ([[28S]]).<ref>{{cite journal | vauthors = Cole JR, Wang Q, Fish JA, Chai B, McGarrell DM, Sun Y, Brown CT, Porras-Alfaro A, Kuske CR, Tiedje JM | display-authors = 6 | title = Ribosomal Database Project: data and tools for high throughput rRNA analysis | journal = Nucleic Acids Research | volume = 42 | issue = Database issue | pages = D633-42 | date = January 2014 | pmid = 24288368 | pmc = 3965039 | doi = 10.1093/nar/gkt1244 }}</ref>

== References ==
{{reflist}}

__FORCETOC__

{{Differentiable computing}}

[[Category:Machine learning]]
[[Category:Bioinformatics]]