Priprava podatkovne zbirke izvencelične mikro RNA v serumu in plazmi

Praznik, Ajda (author), Marc, Janja (mentor), Kranjc, Tilen (co-mentor)

Mikro RNA (miRNA) so endogene, enoverižne RNA, ki uravnavajo izražanje genov in so pomembne pri številnih fizioloških in patoloških procesih (1,2). Kljub temu, da so miRNA v serumu oziroma plazmi privlačne kot obetavni diagnostični biooznačevalci, velik problem predstavlja njihova številčnost, tkivna nespecifičnost in posledično slaba korelacija med profilom mikro RNA in neko določeno boleznijo. Na podlagi sistematičnega pregleda objav o diagnostični uporabnosti miRNA v serumu in plazmi, smo pripravili zbirno tabelo miRNA s pripadajočo diagnostično uporabnostjo, diagnostičnimi parametri kot so specifičnost, občutljivost, natančnost in površina pod ROC krivuljo. Te smo nadalje statistično ovrednotili z metaanalizo, pri čemer smo uporabili programsko opremo Metadisc 1.40. Na osnovi ključnih besed smo oblikovali iskalni profil: ("microRNA"[All Fields] OR "miRNA"[All Fields]) AND ("diagnosis"[All Fields] OR "biomarker"[All Fields]) AND ("serum"[All Fields] OR "plasma"[All Fields]) NOT Review[ptyp] in v njem zajeli 2588 objav. Na osnovi predhodno oblikovanih vključitvenih in izključitvenih kriterijev smo delo nadaljevali s 437 objavami. V teh raziskavah smo preučevali skupno 467 miRNA in njihovih kombinacij. V pripravo ocene diagnostične uporabnosti smo vključili 290 posameznih in 177 kombinacij miRNA. Nadalje smo pridobljene podatke razdelili v štiri podskupine: rak dojke, rak debelega črevesa, nevrodegenerativne bolezni in mišično-skeletne bolezni in jih ločeno ovrednostili z metaanalizo. Metaanaliza podskupin je pokazala, da so skupne vrednosti občutljivosti in specifičnosti, kot merilo učinkovitosti diagnostičnega testa, pri odkrivanju raka dojke znašale 82,8 % in 77,5 %, pri raku debelega črevesja 77,8 % in 78,3 %, pri nevrodegenerativnih boleznih 81,1% in 78,4 % ter pri odkrivanju mišično-skeletnih bolezni 82,4 % in 81,2 %. Površina pod ROC krivuljo, ki klasifikatorje razvršča na osnovi njihovih sposobnosti napovedovanja bolezni, je znašala 0,9108 za napovedovanje raka dojke, 0,8734 za napovedovanje raka debelega črevesja, 0,9015 za napovedovanje nevrodegenerativnih bolezni in 0,8951 pri napovedi mišično-skeletnih bolezni. Posebej natančno smo ovrednotili povezanost mikro RNA za diagnosticiranje Alzheimerjeve bolezeni. MiRNA iz naše podatkovne zbirke smo primerjali s podatki o izražanju miRNA pri bolnikih z Alzheimerjevo boleznijo in pri zdravih kontrolah. S statističnim t testom smo ugotovili, da so štiri miRNA: miRNA-125b(p = 0,05), miRNA-142-3p (p = 0,0002), miRNA-483-3p (p = 0,05) in miRNA-483-5p (p = 0,04) diagnostično učinkovite pri prepoznavanju bolnikov z Alzheimerjevo boleznijo. Na podlagi rezultatov lahko zaključimo, da so miRNA iz podatkovne zbirke diagnostično uporabne. Za namen širše rabe zbirke podatkov smo pripravili tudi spletno aplikacijo, ki bo omogočala lažje iskanje in dostop do podatkov o vrsti ter uporabnosti miRNA v diagnostične namene.

• Language: Slovenian
• Keywords: mikro RNA, biooznačevalci, diagnosticiranje, metaanaliza, sistematičen pregled, podatkovna zbirka
• Work type: Master's thesis/paper
• Organization: FFA - Faculty of Pharmacy
• Year: 2019
• PID: 20.500.12556/RUL-113180
• Publication date in RUL: 11.12.2019

Secondary language

• Language: English
• Title: Design of extracellular micro RNA database in serum and plasma

Abstract

Micro RNA are endogenous, single stranded RNA that regulate gene expression in numerous physiological and pathological processes (1,2). Although, micro RNA are a promising diagnostics biomarkers, the main problem is their abundance, tissue non-specificity and consequently a poor correlation between the miRNA profile and specific disease. Based on systematic reviews of diagnostic performance of serum and plasma miRNA, we designed a database of micro RNA with their diagnostic parameters such as sensitivity, specificity and Area Under the ROC Curve. The data we have obtained from MEDLINE electronic library were processed using the Metadisc 1.40 software. We designed the following search profile: ("microRNA"[All Fields] OR "miRNA"[All Fields]) AND ("diagnosis"[All Fields] OR "biomarker"[All Fields]) AND ("serum"[All Fields] OR "plasma"[All Fields]) NOT Review[ptyp]. We found 2588 publications. Based on previously formed inclusion and exclusion criteria, we continued our work with 437 publications. A total of 467 miRNAs and their combinations were included. In assessment of the diagnostic purpose we included 290 individual and 177 miRNA combinations. The data was then divided into four subgroups: breast cancer, colorectal cancer, neurodegenerative diseases and musculoskeletal diseases. Pooled sensitivity and specificity assessing the effectiveness of diagnostic performance were: 82.8% in 77.5% for breast cancer, 77.8% and 78.3% for colorectal cancer, 81.1% and 78.4% for neurodegenerative diseases and 82.4% and 81.2% for musculoskeletal diseases. The Area Under the ROC Curve, which classifies development based on major disease prediction abilities, was 0.9108 for breast cancer, 0.8734 for colorectal cancer, 0.9015 for neurodegenerative diseases and 0.8951 for musculoskeletal diseases. Furthermore, we evaluated miRNA as biomarkers for Alzheimer's disease. We compared miRNA from our database with the miRNA expression data in patients with Alzheimer's disease and in healthy controls. With statistical t test, we confirmed that four miRNA: miRNA-125b (p = 0,05), miRNA-142-3p (p = 0,0002), miRNA-483-3p (p = 0,05) and miRNA-483-5p (p = 0,04), were diagnostically effective in identifying patients with Alzheimer's disease. We concluded that miRNA in our database could be used as diagnostic biomarkers for our four disease groups. For the purpose of a wider use of the database, we also created an online application that will ease searching and access to the information regarding usefulness of miRNA in diagnostics purposes.

• Keywords: micro RNA, biomarkers, diagnostics, meta-analysis, systematic review, database