The “MetaCopepod” project: Designing an integrated DNA metabarcoding and image analysis approach to study and monitor biodiversity of zooplanktonic copepods. PanagiotisPanagiotis KasapidisKasapidis HellenicHellenic CentreCentre forfor MarineMarine ResearchResearch (HCMR),(HCMR), InstituteInstitute ofof MarineMarine Biology,Biology, BiotechnologyBiotechnology andand AquacultureAquaculture P.O.Box 2214, 71003 Heraklion Crete, Greece 1 The “MetaCopepod” project Aim: to develop a novel methodology, based on the combination of DNA metabarcoding and image analysis, to assess and monitor the diversity of marine zooplanktonic copepods (and cladocera), in the Mediterranean and the Black Sea, in a high-throughput, cost- effective, accurate and quantitative way. Coordinator: Dr. Panagiotis Kasapidis, Hellenic Centre for Marine Research (HCMR), GREECE Study area: Mediterranean and the Black Sea Duration: Feb. 2014 – Oct. 2015 Budget: 180,000 euros Funding: European Social Fund (ESF) and National Funds through the National Strategic Reference Framework (NSRF) 2007-2013, Operational Programme "Education and Life-Long Learning", Action "ARISTEIA II", Greek Ministry of Education and Religious Affairs, General Secretary of Research and Technology. 2 Studying zooplankton diversity: limitations of traditional approaches ● Quite laborious (sorting, identification under stereoscope) → bottleneck in sample processing. ● Requires local taxonomic expertise ● Difficult to identify immature stages ● Misidentifications ● Cryptic species 3 Image analysis + Pros High throughput Quantitative results (abundance, size spectra, biomass) - Cons Low resolution → can recognize taxa similar to the ability of a trained taxonomists to identify a taxon under stereoscope at a glance “train” image analysis software separately for different types of zooplankton communities (not once for all). 4 TGTTCA 5 NGS AGTTCT AGATCT AGTTCA bulksample AGTTCA requires a curatedwell and studying marine biodiversity marine studying DNA metabarcoding (NGS analysis) for for analysis) (NGS metabarcoding DNA Biases in species' relative abundance mostly due to PCR Results are semiquantitative due to the PCR No No need of taxonomy expert but Potentially higher accuracy in species' identification (even Faster processing of the samples amplification bias - Cons: complete reference genetic database for difficult fortaxa, difficult immature stages, cryptic species) + Pros: bothmethods inorderto increase accuracy inassessing zooplanktondiversity. The“MetaCopepod” project toaims combine the Prosof Workflow of the “MetaCopepod” project WP7: Dissemination WP1 sampling Sampling and imaging morphological identification of pseudo-samples bulk samples image acquisition of bulk copepod species samples image acquisition of training of image WP2 individuals image analysis analysis software Image analysis DNA extraction WP3 WP4 DNA extraction in silico primer design DNA amplification with PCR amplification for standard COI barcode and testing DNA barcoding- and selected barcode selected barcode primers DNA Metabarcoding genetic reference DB Sanger sequencing High throughput next generation DNA sequencing online management system phylogenetic analysis local reference database raw sequencing data WP5 Bioinformatic Construction of Bioinformatic analysis DNA based species comparison with image pipeline bioinformatic pipeline identification analysis data WP6: Demonstration of methodology (Copepod biomonitoring) 6 Sampling Napoli Saronikos Annaba Cretan Sea Partner Network and sampling stations. Monthly sampling ( ), non-regular sampling ( ) 7 Samples 108 zooplankton samples collected and preserved in 95% EtOH 80 zooplankton samples were scanned for image analysis and then used for DNA metabarcoding 97 copepod species were morphologically identified ( >350 specimen collections from different zooplankton samples) For standardizing the methodology: Five of the zooplankton samples were taxonomically identified by taxonomist Six pseudosamples (mock samples) created by mixing taxonomically identified specimens at known numbers (6-16 species, 1-3 individuals per species) 8 Image analysis (Epson scanner similar to Zooscan and ZooImage software) For “training” the image analysis software we used: scanned images of taxonomically identified taxa (“gold” standards) taxonomically identified taxa from scanned zooplankton images (“silver” standards) The image analysis software was evaluated by: self-evaluation using pseudosamples and taxonomically identified samples 9 Self evaluation of the Image Analysis software Copepods Cladocera Bad focus or multiple Other zooplankton Other (artifacts) Mostly correct identification, but confusion in certain taxa (e.g. Clausocalanus and Ctenocalanus) 10 ● ● IA software performs well both for taxa recognition and abundance estimation abundance and recognition for taxa both well performs software IA IA by and bya taxonomist both Gulf analyzed from Saronikos samples Three Evaluation of the Evaluation Image Analysis software Number of individuals with taxonomically taxonomically with samples identified TAX IA TAX SG1-0214 SG1-0214_1 TAX IA TAX SG1-0214_2 TAX IA TAX 11 Larger Categories for Image categories Analysis Taxa in each category Calanus helgolandicus Calanus_helgolandicus Candacia Candacia-Paracandacia Paracandacia simplex Big Calanoida Euchirella Euchirella Nannocalanus minor Nannocalanus minor Neocalanus Neocalanus Pleuromamma Pleuromamma Acartia Acartiidae Paracartia grani Centropagidae Centropages Medium Calanoida Categories identified with Temora accuracy by Image analysis Temora Aetidus Lucicutia (standardized for Saronikos Calocalanus Gulf) Clausocalanus Paracalanus Ctenocalanus vanus Small Calanoida Small Calanoida Phaenna spinifera Scolecithricella & Scolecithrix Isias clavipes Mecynocera clausi Cyclopoida Oithona Oithona Euterpina acutifrons Euterpina acutifrons Harpacticoida Clytemnestra Clytemnestra other Harpacticoida n.d. Micro & Macrosettela Micro & Macrosettela Coryceaus Coryceaidae Farranula rostrata Poecilostomatoida Oncaea Oncaeidae Triconia Ctenopoda Penilia_avirostris Penilia avirostris Evadne Pseudevadne Onychopoda Onychopoda Podon Pleopis polyphemoides 12 Critical factors for DNA metabarcoding Primers designed on conserved regions across target taxa→ reduce amplification bias amplify a variable region → high resolution, ideally at species' level amplify a short region→ easier PCR amplification even for degraded samples Genetic reference database – well curated (no errors or misidentifications) – as complete as possible 13 Primer design and reference database Primers were designed (ecoPrimer software), which amplify a region of ~150 bp of the mitochondrial 16S rRNA gene. Reference database (GenBank largely incomplete for 16S rRNA) Out of the 97 taxa taxonomically identified 93 lacked 16S sequences and 36 lacked COI sequences in GenBank We performed DNA barcoding for both COI and 16S genes 50 new additions of 16S barcode (55 species sequenced in total) 8 new additions for standard COI barcode (48 species sequenced in total) Sapphirina opalina 14 Evaluation of the 16S barcode to discriminate species CopF barcode (~120 bp) CopR-Un Alignment of 16S copepod sequences from NCBI and from MetaCopepod (produced with universal primers) 15 Median-Joining network for the 16S rRNA barcode for copepod species 16 DNA metabarcoding of zooplankton samples Unsorted zooplankton samples, taxonomically identified samples and pseudosamples → DNA extraction, PCR and sequencing on a MiSeq Illumina platform. Raw sequencing data analyzed using a bioinformatic pipeline constructed for the project. Sequences assigned to taxa using the reference database 17 necessary. checks more technical actualhandling?→ to Systematic?Due data. genetic and between speciesfew (e.g. For biomass)]. [ Corycaeidae the exceptfor total sample biomass), the All speciesof by analysis as genetic low 0.12% (even detected speciesindividuals of each. 1-3 Pseudosamplescontained 6-16 relative abundance in the pseudosample 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0 Acartia clausi-PS1 Acartia clausi-PS3 comparison between genetic Pseudosamples: Acartia clausi-PS4 Acartia clausi-PS5 Agetus typicus-PS2 data, % abundance (%N) and % biomass and data, % abundance (%N) Calanus helgolandicus-PS2 Calanus helgolandicus-PS3 Calocalanus styliremis-PS6 Calanus helgolandicus Calanus Candacia bispinosa-PS2 Candacia simplex-PS2 Centropages ponticus-PS4 Centropages typicus-PS1 Farranula rostrata rostrata Farranula Centropages typicus-PS3 Centropages typicus-PS5 Clausocalanus arcuicornis-PS1 Clausocalanus furcatus-PS2 Clausocalanus furcatus-PS4 Clausocalanus furcatus-PS6 Clausocalanus jobei-PS1 Clausocalanus lividus-PS2 Clausocalanus lividus-PS3 Clausocalanus paululus-PS6 and and Ctenocalanus vanus-PS1 (max 6.5% biomass) and biomass) 6.5% (max Euchirella rostrata-PS2 Euchirella rostrata-PS3 Euterpina acutifrons Euterpina Euterpina acutifrons-PS5 Evadne nordmanni-PS1 Farranula rostrata-PS2 Farranula rostrata-PS6 Haloptilus longicornis-PS6 Nannocalanus minor-PS2 Nannocalanus minor-PS3 Neocalanus gracilis-PS2 Neocalanus gracilis-PS3 Oncaea media-PS1 Oncaea media-PS2 Oncaea scottodicarloi-PS1 ), large discrepancies large ), Oncaea scottodicarloi-PS2 Agetus typicus Agetus Oncea scottodicarloi-PS5 Paracalanus denudatus-PS6 Paracalanus nanus-PS2 Paracalanus parvus-PS1 Paracalanus parvus-PS4 Paracalanus parvus-PS5 %Genetic %Biomass %N Penilia avirostris-PS4 Pleuromamma abdominalis-PS2 Pleuromamma gracilis-PS2 Podon