Forest soil bacterial community analysis using high-throughput amplicon sequencing
Kuupäev
2017-10-10
Autorid
Ajakirja pealkiri
Ajakirja ISSN
Köite pealkiri
Kirjastaja
Abstrakt
Muldade rikkalike mikroobikoosluste uurimist on siiani palju takistanud tõsiasi, et enamik mulla mikroobe on kultiveerimatud. Seda kitsaskohta aitab leevendada lähenemine nimega metagenoomika, mis tähistab uurimistööd otse keskkonnaproovidest eraldatud geneetilise materjaliga.
Selliste andmete kasutamiseks on levinud meetodid, mille abil grupeeritakse (klasterdatakse) kogutud DNA järjestused ad-hoc taksonoomilistesse üksustesse nn. OTU-desse (Operational Taxonomic Unit). Nii võib OTU-desse klasterdatud järjestusi kasutades hinnata bakterikoosluste mitmekesisust ja liigilist koostist. Saadud OTU-de arvukuse numbreid annab kasutada mitmesugustes erinevates analüüsides kui asendajaid tavapärasematele taksonoomilistele üksustele. Niisama kiire, kui on olnud uute sekveneerimistehnoloogiate areng, on ka olnud uute tööriistade arvu kasv – viimase kümnendi jooksul on loodud hulk programme, mis on mõeldud eelpoolmainitud OTU-de moodustamiseks DNA järjestuste andmetest.
Antud doktoritöö töö keskendub sellele, kuidas mõjutavad erinevad OTU loomise meetodid edasisi analüüse ning järeldusi.
Selleks kasutati järjestusandmeid artiklist “Bacterial community structure and its relationship to soil physico-chemical characteristics in alder stands with different management histories” ning erinevaid OTU klasterdamise meetodeid. OTU-d loodi erinevate programmide abil (Mothur,CROP,UCLUST,Swarm) – seejärel viidi läbi koosluste mitmesugused statistilised analüüsid.
OTU andmete analüüs andis üldjoontes samasuguseid tulemusi. Seda visualiseerivad hästi töös olevad joonised. OTU arvude ja mitmekesisusindeksi statistilised testid ei leidnud statistiliselt olulist erinevust eri klasterdusmeetodite vahel.
Kasutatud klasterdamismeetoditest jäid parimaina silma paistma CROP ja UCLUST meetodid.Lisaks näitasid analüüsid ka osade statistiliste meetodite eeliseid teiste ees sedasorti OTU andmete käsitlemisel
The soil as a central agent in many ecological processes has received a lot of research attention from many different angles. The investigation of the rich microbiome of the soil has been slowed by the fact that most of the microbes are unculturable. This gap can be filled by the metagenomics which is a field that deals with genetic material directly acquired form environmental samples. The analysis of 16S rDNA data usually begins with the construction of operational taxonomicunits (OTUs): clusters of reads that differ by less than a fixed sequence dissimilarity threshold. Consequently, the obtained sample-by-OTU abundance table serves as the basis for further statistical and exploratory analysis. During the last decade, a plethora of tools based on different principles and having different computational requirements to perform aforementioned OTU clustering has been created. This work we take an interest in the differences of the final outcome of series of analyses when different OTU clustering methods are used and also have a comparision of these methods. We used the dataset published in “Bacterial community structure and its relationship to soil physico-chemical characteristics in alder stands with different management histories” and analysed it using different software packages for processing bioinformatics data: Mothur UCLUST, CROP, Swarm. The results of analyses were on the whole quite similar and comparable.The differences between OTU numbers and diversity indeces were statistically not significant. The CROP and UCLUST methods stood out by their quality and useability. The work also showed the practicality of robust statistical methods when working with OTU data.
The soil as a central agent in many ecological processes has received a lot of research attention from many different angles. The investigation of the rich microbiome of the soil has been slowed by the fact that most of the microbes are unculturable. This gap can be filled by the metagenomics which is a field that deals with genetic material directly acquired form environmental samples. The analysis of 16S rDNA data usually begins with the construction of operational taxonomicunits (OTUs): clusters of reads that differ by less than a fixed sequence dissimilarity threshold. Consequently, the obtained sample-by-OTU abundance table serves as the basis for further statistical and exploratory analysis. During the last decade, a plethora of tools based on different principles and having different computational requirements to perform aforementioned OTU clustering has been created. This work we take an interest in the differences of the final outcome of series of analyses when different OTU clustering methods are used and also have a comparision of these methods. We used the dataset published in “Bacterial community structure and its relationship to soil physico-chemical characteristics in alder stands with different management histories” and analysed it using different software packages for processing bioinformatics data: Mothur UCLUST, CROP, Swarm. The results of analyses were on the whole quite similar and comparable.The differences between OTU numbers and diversity indeces were statistically not significant. The CROP and UCLUST methods stood out by their quality and useability. The work also showed the practicality of robust statistical methods when working with OTU data.
Kirjeldus
Väitekirja elektrooniline versioon ei sisalda publikatsioone
Märksõnad
metsamullad, mullamikroobid, mikroobikooslused, klasteranalüüs, forest soils, soil microbes, microbial communities, cluster analysis