On the Relation Between Linear Autoencoders and Non-Negative Matrix Factorization for Mutational Signature Extraction
Since its introduction, non-negative matrix factorization (NMF) has been a popular tool for extracting interpretable, low-dimensional representations of high-dimensional data. However, several recent studies have proposed replacing NMF with autoencoders. The increasing popularity of autoencoders war...
Uložené v:
| Vydané v: | Journal of computational biology Ročník 32; číslo 5; s. 461 |
|---|---|
| Hlavní autori: | , , , , |
| Médium: | Journal Article |
| Jazyk: | English |
| Vydavateľské údaje: |
United States
01.05.2025
|
| Predmet: | |
| ISSN: | 1557-8666, 1557-8666 |
| On-line prístup: | Zistit podrobnosti o prístupe |
| Tagy: |
Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
|
| Abstract | Since its introduction, non-negative matrix factorization (NMF) has been a popular tool for extracting interpretable, low-dimensional representations of high-dimensional data. However, several recent studies have proposed replacing NMF with autoencoders. The increasing popularity of autoencoders warrants an investigation on whether this replacement is in general valid and reasonable. Moreover, the exact relationship between non-negative autoencoders and NMF has not been thoroughly explored. Thus, a main aim of this study is to investigate in detail the relationship between autoencoders and NMF. We define a non-negative linear autoencoder, AE-NMF, which is mathematically equivalent with convex NMF, a constrained version of NMF. The performance of NMF and the non-negative linear autoencoder is compared within the context of mutational signature extraction from simulated and real-world cancer genomics data. We find that the reconstructions based on NMF are more accurate compared with AE-NMF, while the signatures extracted using both methods exhibit comparable consistency and performance when externally validated. These findings suggest that AE-NMF, the linear non-negative autoencoders investigated in this article, do not provide an improvement of NMF in the field of mutational signature extraction. Our study serves as a foundation for understanding the theoretical implication of replacing NMF with non-negative autoencoders. |
|---|---|
| AbstractList | Since its introduction, non-negative matrix factorization (NMF) has been a popular tool for extracting interpretable, low-dimensional representations of high-dimensional data. However, several recent studies have proposed replacing NMF with autoencoders. The increasing popularity of autoencoders warrants an investigation on whether this replacement is in general valid and reasonable. Moreover, the exact relationship between non-negative autoencoders and NMF has not been thoroughly explored. Thus, a main aim of this study is to investigate in detail the relationship between autoencoders and NMF. We define a non-negative linear autoencoder, AE-NMF, which is mathematically equivalent with convex NMF, a constrained version of NMF. The performance of NMF and the non-negative linear autoencoder is compared within the context of mutational signature extraction from simulated and real-world cancer genomics data. We find that the reconstructions based on NMF are more accurate compared with AE-NMF, while the signatures extracted using both methods exhibit comparable consistency and performance when externally validated. These findings suggest that AE-NMF, the linear non-negative autoencoders investigated in this article, do not provide an improvement of NMF in the field of mutational signature extraction. Our study serves as a foundation for understanding the theoretical implication of replacing NMF with non-negative autoencoders.Since its introduction, non-negative matrix factorization (NMF) has been a popular tool for extracting interpretable, low-dimensional representations of high-dimensional data. However, several recent studies have proposed replacing NMF with autoencoders. The increasing popularity of autoencoders warrants an investigation on whether this replacement is in general valid and reasonable. Moreover, the exact relationship between non-negative autoencoders and NMF has not been thoroughly explored. Thus, a main aim of this study is to investigate in detail the relationship between autoencoders and NMF. We define a non-negative linear autoencoder, AE-NMF, which is mathematically equivalent with convex NMF, a constrained version of NMF. The performance of NMF and the non-negative linear autoencoder is compared within the context of mutational signature extraction from simulated and real-world cancer genomics data. We find that the reconstructions based on NMF are more accurate compared with AE-NMF, while the signatures extracted using both methods exhibit comparable consistency and performance when externally validated. These findings suggest that AE-NMF, the linear non-negative autoencoders investigated in this article, do not provide an improvement of NMF in the field of mutational signature extraction. Our study serves as a foundation for understanding the theoretical implication of replacing NMF with non-negative autoencoders. Since its introduction, non-negative matrix factorization (NMF) has been a popular tool for extracting interpretable, low-dimensional representations of high-dimensional data. However, several recent studies have proposed replacing NMF with autoencoders. The increasing popularity of autoencoders warrants an investigation on whether this replacement is in general valid and reasonable. Moreover, the exact relationship between non-negative autoencoders and NMF has not been thoroughly explored. Thus, a main aim of this study is to investigate in detail the relationship between autoencoders and NMF. We define a non-negative linear autoencoder, AE-NMF, which is mathematically equivalent with convex NMF, a constrained version of NMF. The performance of NMF and the non-negative linear autoencoder is compared within the context of mutational signature extraction from simulated and real-world cancer genomics data. We find that the reconstructions based on NMF are more accurate compared with AE-NMF, while the signatures extracted using both methods exhibit comparable consistency and performance when externally validated. These findings suggest that AE-NMF, the linear non-negative autoencoders investigated in this article, do not provide an improvement of NMF in the field of mutational signature extraction. Our study serves as a foundation for understanding the theoretical implication of replacing NMF with non-negative autoencoders. |
| Author | Brøndum, Rasmus Froberg Hobolth, Asger Egendal, Ida Pelizzola, Marta Bøgsted, Martin |
| Author_xml | – sequence: 1 givenname: Ida orcidid: 0000-0002-6189-6053 surname: Egendal fullname: Egendal, Ida organization: Clinical Cancer Research Center, Aalborg University Hospital, Aalborg, Denmark – sequence: 2 givenname: Rasmus Froberg surname: Brøndum fullname: Brøndum, Rasmus Froberg organization: Clinical Cancer Research Center, Aalborg University Hospital, Aalborg, Denmark – sequence: 3 givenname: Marta surname: Pelizzola fullname: Pelizzola, Marta organization: Department of Mathematics, Aarhus University, Aarhus, Denmark – sequence: 4 givenname: Asger surname: Hobolth fullname: Hobolth, Asger organization: Department of Mathematics, Aarhus University, Aarhus, Denmark – sequence: 5 givenname: Martin surname: Bøgsted fullname: Bøgsted, Martin organization: Clinical Cancer Research Center, Aalborg University Hospital, Aalborg, Denmark |
| BackLink | https://www.ncbi.nlm.nih.gov/pubmed/40113251$$D View this record in MEDLINE/PubMed |
| BookMark | eNpNkMtOwzAQRS1URB-wZIu8ZJNiO3bsLEvVAlIfEo91ZCeTEpTaxXGg8PVEtEis5s7VmbOYIepZZwGhS0rGlKj0Jt-aMSOMj4lU_AQNqBAyUkmS9P7lPho2zRshNE6IPEN9TiiNmaAD1K4tDq-AH6HWoXIW30L4BLB4UVnQHk_a4MDmrgDfYG0LvHI2WsGmgz8AL3Xw1R7PdR6cr74PhtJ5vGzD76Jr_FRtrA6tBzzbB9-RXX2OTktdN3BxnCP0Mp89T--jxfruYTpZRHkseYhEClIpalIlwbDYcKVTnhqiSMEKIhNppM4TnjMhNYupLApBZMx5WSacmtiwEbo-eHfevbfQhGxbNTnUtbbg2ibrblIlqBSyQ6-OaGu2UGQ7X221_8r-XsV-AOnUbXI |
| CitedBy_id | crossref_primary_10_1016_j_cosrev_2025_100788 |
| ContentType | Journal Article |
| DBID | CGR CUY CVF ECM EIF NPM 7X8 |
| DOI | 10.1089/cmb.2024.0784 |
| DatabaseName | Medline MEDLINE MEDLINE (Ovid) MEDLINE MEDLINE PubMed MEDLINE - Academic |
| DatabaseTitle | MEDLINE Medline Complete MEDLINE with Full Text PubMed MEDLINE (Ovid) MEDLINE - Academic |
| DatabaseTitleList | MEDLINE - Academic MEDLINE |
| Database_xml | – sequence: 1 dbid: NPM name: PubMed url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed sourceTypes: Index Database – sequence: 2 dbid: 7X8 name: MEDLINE - Academic url: https://search.proquest.com/medline sourceTypes: Aggregation Database |
| DeliveryMethod | no_fulltext_linktorsrc |
| Discipline | Biology Mathematics |
| EISSN | 1557-8666 |
| ExternalDocumentID | 40113251 |
| Genre | Journal Article |
| GroupedDBID | --- 0R~ 29K 34G 39C 4.4 53G 5GY ABBKN ABEFU ACGFO ADBBV AENEX AFOSN AI. ALMA_UNASSIGNED_HOLDINGS BAWUL BNQNF CAG CGR COF CS3 CUY CVF D-I DIK DU5 EBS ECM EIF EJD F5P IAO IER IGS IHR IM4 ITC MV1 NPM NQHIM O9- P2P R.V RIG RML RMSOB RNS TN5 TR2 UE5 VH1 7X8 SCNPE |
| ID | FETCH-LOGICAL-c374t-59e7881b987eb23b48a949b080d2d0767b7ac64c257a2317dd507344ff641b3b2 |
| IEDL.DBID | 7X8 |
| ISICitedReferencesCount | 1 |
| ISICitedReferencesURI | http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=001448234000001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| ISSN | 1557-8666 |
| IngestDate | Fri Sep 05 14:34:30 EDT 2025 Tue May 13 01:30:45 EDT 2025 |
| IsPeerReviewed | true |
| IsScholarly | true |
| Issue | 5 |
| Keywords | convex non-negative matrix factorization non-negative matrix factorization mutational signatures non-negative autoencoders |
| Language | English |
| LinkModel | DirectLink |
| MergedId | FETCHMERGED-LOGICAL-c374t-59e7881b987eb23b48a949b080d2d0767b7ac64c257a2317dd507344ff641b3b2 |
| Notes | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 23 |
| ORCID | 0000-0002-6189-6053 |
| PMID | 40113251 |
| PQID | 3179851757 |
| PQPubID | 23479 |
| ParticipantIDs | proquest_miscellaneous_3179851757 pubmed_primary_40113251 |
| PublicationCentury | 2000 |
| PublicationDate | 2025-05-01 |
| PublicationDateYYYYMMDD | 2025-05-01 |
| PublicationDate_xml | – month: 05 year: 2025 text: 2025-05-01 day: 01 |
| PublicationDecade | 2020 |
| PublicationPlace | United States |
| PublicationPlace_xml | – name: United States |
| PublicationTitle | Journal of computational biology |
| PublicationTitleAlternate | J Comput Biol |
| PublicationYear | 2025 |
| SSID | ssj0013607 |
| Score | 2.4395401 |
| Snippet | Since its introduction, non-negative matrix factorization (NMF) has been a popular tool for extracting interpretable, low-dimensional representations of... |
| SourceID | proquest pubmed |
| SourceType | Aggregation Database Index Database |
| StartPage | 461 |
| SubjectTerms | Algorithms Autoencoder Computational Biology - methods Genomics - methods Humans Mutation Neoplasms - genetics |
| Title | On the Relation Between Linear Autoencoders and Non-Negative Matrix Factorization for Mutational Signature Extraction |
| URI | https://www.ncbi.nlm.nih.gov/pubmed/40113251 https://www.proquest.com/docview/3179851757 |
| Volume | 32 |
| WOSCitedRecordID | wos001448234000001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| hasFullText | |
| inHoldings | 1 |
| isFullTextHit | |
| isPrint | |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV1JT4NAFJ6o1UQPLnWrW8bE61g6DAxzMtW08SDYRE16IzAzND0ItYCp_943QNWLiYkXEg4QMrzle9v3ELoCp8-tSHLCOHMJ-OuEeFLFRLlcaiqZoDXP7AMPAm88FqMm4ZY3bZVLm1gZapVJkyPv2oZZywFnx29mb8RsjTLV1WaFxipq2QBljFTz8Y8qgluNS4PLBEsMOL3h2LQ80ZWvMQSHlF2Di2S_o8vKywx3_vt9u2i7wZe4XwvEHlrRaRtt1BsnP9poy_-iac33UfmYYrjFy444fFt3bWGIUEEDcL8sMkN0aZqdcZQqHGQpCfSkIgvHvmH3X-BhtbGnGefEgIGxXxZNihE_TSc1dSgeLIp5PURxgF6Gg-e7e9LsYSDS5qwgjtCGdD4WHoc43I6ZFwkmYsCaiiqLuzzmkXSZBO2PAC5ypUAAbMaSxGW92I7pIVpLs1QfI-wKqrQneomjFaMK3skNY56yPJUwQEYddLk83RDk3BQvolRnZR5-n28HHdW_KJzVhBwhxIgQVDu9kz88fYo2qVnhW_UsnqFWAlquz9G6fC-m-fyiEiC4BiP_E1lL0Fg |
| linkProvider | ProQuest |
| openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=On+the+Relation+Between+Linear+Autoencoders+and+Non-Negative+Matrix+Factorization+for+Mutational+Signature+Extraction&rft.jtitle=Journal+of+computational+biology&rft.au=Egendal%2C+Ida&rft.au=Br%C3%B8ndum%2C+Rasmus+Froberg&rft.au=Pelizzola%2C+Marta&rft.au=Hobolth%2C+Asger&rft.date=2025-05-01&rft.issn=1557-8666&rft.eissn=1557-8666&rft_id=info:doi/10.1089%2Fcmb.2024.0784&rft.externalDBID=NO_FULL_TEXT |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1557-8666&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1557-8666&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1557-8666&client=summon |