prikaz prve stranice dokumenta Machine learning assisted determination of linearly independent set of generalized molecular coordinates
  Download
PDF 2.35 MB
master's thesis
Machine learning assisted determination of linearly independent set of generalized molecular coordinates
Zagreb: University of Zagreb, Faculty of Science, 2021. urn:nbn:hr:217:138103
Ostojić, Tea
University of Zagreb
Faculty of Science
Department of Chemistry

Institutional repository: Repository of the Faculty of Science

Cite this document

Ostojić, T. (2021). Machine learning assisted determination of linearly independent set of generalized molecular coordinates (Master's thesis). Zagreb: University of Zagreb, Faculty of Science. Retrieved from https://urn.nsk.hr/urn:nbn:hr:217:138103

Ostojić, Tea. "Machine learning assisted determination of linearly independent set of generalized molecular coordinates." Master's thesis, University of Zagreb, Faculty of Science, 2021. https://urn.nsk.hr/urn:nbn:hr:217:138103

Ostojić, Tea. "Machine learning assisted determination of linearly independent set of generalized molecular coordinates." Master's thesis, University of Zagreb, Faculty of Science, 2021. https://urn.nsk.hr/urn:nbn:hr:217:138103

Ostojić, T. (2021). 'Machine learning assisted determination of linearly independent set of generalized molecular coordinates', Master's thesis, University of Zagreb, Faculty of Science, accessed 17 May 2024, https://urn.nsk.hr/urn:nbn:hr:217:138103

Ostojić T. Machine learning assisted determination of linearly independent set of generalized molecular coordinates [Master's thesis]. Zagreb: University of Zagreb, Faculty of Science; 2021 [cited 2024 May 17] Available at: https://urn.nsk.hr/urn:nbn:hr:217:138103

T. Ostojić, "Machine learning assisted determination of linearly independent set of generalized molecular coordinates", Master's thesis, University of Zagreb, Faculty of Science, Zagreb, 2021. Available at: https://urn.nsk.hr/urn:nbn:hr:217:138103

Metadata
TitleMachine learning assisted determination of linearly independent set of generalized molecular coordinates
Title (croatian)Određivanje linearno nezavisnog skupa generaliziranih molekularnih koordinata primjenom strojnog učenja
AuthorTea Ostojić
MentorTomica Hrenar (mentor)
Committee memberTomica Hrenar (predsjednik povjerenstva)
Committee memberIvan Kodrin (član povjerenstva)
Committee memberŽeljka Soldin (član povjerenstva)
Committee memberBranimir Bertoša (član povjerenstva)
GranterUniversity of Zagreb
Faculty of Science
(Department of Chemistry)
Zagreb
Defense date and country2021-05-18, Croatia
Scientific / art field,
discipline and subdiscipline		NATURAL SCIENCES
Chemistry
Abstract
Trajectory analysis can provide information on the complete conformational or configurational space of molecular systems and their reactivity. Numerical analysis of trajectories can be performed in any type of coordinate system, e.g. Cartesian coordinate system, internal or normal coordinates, etc. In some applications, various coordinate systems will hold redundant information due to the linear dependence. A set of molecular coordinates contains null-vectors that introduce numerical instabilities and noise in the processing. In the case of a larger number of defined coordinates, there is also a problem with storage and data processing. In this work, the procedure for constructing generalized and linearly independent set of internal molecular coordinates will be evaluated. Generalized internal coordinates are chemically intuitive and can be defined in a number of ways, for example: distances, angles or dihedral angles. By applying machine learning to molecular dynamics trajectory, a complete and linearly dependent set of generalized internal coordinate distances will be reduced to the linearly independent set that still contains all relevant information about conformational or configurational space, reactivity and molecular motion. Given procedure will be applied to the molecule of (R)-cinchonidine.
Abstract (croatian)
Analiza trajektorije može pružiti informacije o potpunom konformacijskom ili konfiguracijskom prostoru molekularnih sustava i njihovoj reaktivnosti. Numerička analiza trajektorija može se provesti u bilo kojem koordinatnom sustavu, npr. Cartesiusov koordinatni sustav, sustav internih ili normalnih koordinata, itd. No, u nekim aplikacijama različiti koordinatni sustavi zadržavaju suvišne informacije zbog linearne zavisnosti. Taj skup molekularnih koordinata sadrži nul-vektore koji onda uzrokuju numeričke nestabilnosti i numerički šum. U slučaju većeg broja definiranih koordinata također postoji i problem s pohranom te obradom podataka. U ovom će se radu razraditi postupak za kreiranje generaliziranog i linearno nezavisnog skupa internih koordinata. Generalizirane interne koordinate kemijski su intuitivne, a mogu se definirati na više načina, na primjer: međuatomne udaljenosti, kutovi ili diedarski kutovi. Primjenom strojnog učenja na trajektoriju molekularne dinamike, potpun i linearno zavisan skup generaliziranih internih koordinata međuatomnih udaljenosti reducirat će se na linearno nezavisan skup, ali skup koji još uvijek sadrži sve relevantne informacije o konformacijskom ili konfiguracijskom prostoru, reaktivnosti i gibanju molekula. Razrađeni postupak primijenit će se na molekulu (R)-cinhonidina.
Keywords
Keywords (english)
Languageenglish
URN:NBNurn:nbn:hr:217:138103
ProjectNumber: IP-2016-06-3775
Title: Aktivnošću i in silico usmjeren dizajn malih bioaktivnih molekula
Title: Activity and in silico guided design of bioactive small molecules
Acronym: ADESIRE
Leader: Tomica Hrenar
Jurisdiction: Croatia
Funder: HRZZ
Funding stream: IP
Study programmeTitle: Graduate university studies in chemistry; specializations in: Research programme
Course: Research programme
Study programme type: university
Study level: graduate
Academic / professional title: magistar/magistra kemije (magistar/magistra kemije)
Type of resourceText
File originBorn digital
Access conditionsOpen access
Embargo expiration date: 2023-05-18
Terms of useLicence In copyright icon
Created on2021-05-28 12:40:46