SPAAM2 - Programme

Session Structure

Length: 2 days

All times CEST (see here for Time Zone conversion)

Day 1

Day 2

Abstracts

All sessions will include ‘icebreaker’ talks by members of the SPAAM community, who will present on their own experiences and challenges working ancient metagenomics to serve as starting point for discussions.

:warning: All titles and abstracts are subject to minor changes!

Session 1

Title Trash in, trash out: Optimizing and standardizing laboratory practices in ancient metagenomics

Chair Irina Velsko

Abstract Optimizing laboratory practices is necessary to reduce the amount of contaminating DNA in a sample extract, and subsequently in sequencing data. All sequenced contaminating DNA takes up information that should be for sample DNA, reducing the efficiency of workflows, both financially and bioinformatically. Further, sample biomass influences the amount of “spill-over” that is observed in sequence data, where the contents of higher biomass samples may show up in lower biomass samples. In this session we will discuss how best to prevent or reduce introduction of contamination to sample extraction and library building protocols, as well as limiting sample cross-contamination. These techniques aim to improve sequencing efficiency by reducing the amount of sequenced contamination and increasing the amount of sequenced sample DNA.

Icebreaker presenters Jaelle Brealey and Zandra Fagernäs

Icebreaker mini-abstract Reducing contamination: how to minimize the amount of trash Contaminants can be introduced to samples during excavation, post-excavation handling and storage and processing in the laboratory. Minor differences between sample processing batches and cross-contamination between samples processed together can also cause problems in downstream analyses. In this talk we will discuss methods to reduce and control for external contamination, batch effects and cross-contamination during laboratory processing of ancient samples.

Session 2

Title Removing the persistent trash: Challenges in genotyping and filtering out contaminant reads from genome alignments

Chair Åshild (Ash) Vågene

Abstract Genotyping is the characterization of variants (SNPs, Indels) that occur at the single genome level. Accurate genotyping of ancient microbial genomes is crucial for downstream analyses, such as phylogenetic placement, SNP effect analysis and molecular dating. Genotyping is also important for taxonomic profiling when trying to perform strain separation. Thus, inaccuracies in genotyping/variant calling can have an enormous impact on the interpretation of the results.

Confounding factors related to these genomes being a) ancient, b) from metagenomic backgrounds, can negatively impact genotyping efforts by causing multiallelic variant calls that infringe on thresholds for calling homozygous variants. These factors include ancient DNA damage, short reads prone to mis-mapping in regions of genomic rearrangement and cross-mapping of contaminating sequences derived from genetically similar organisms from the specimens’ burial/storage environment. Approaches such as edit distance and analysing the allele frequency of multiallelic positions are commonly used to give an indication of the level of ‘contamination’ in an alignment. Both molecular and computational approaches exist to mitigate the effects of contaminant and mis-mapping reads in genotype calling.

In this session we will discuss issues and problems relating to genotyping of ancient microbial single-genomes, with emphasis on (but not restricted to) haploid bacterial and viral genomes. The aim of the session is not only to make others aware of specific issues/scenarios that may arise from working with different types of microbes, but also to have a general discussion of the best approaches to tackle such problems.

Icebreaker presenters Susanna Sabin and Kun Huang

Icebreaker mini-abstract 1 One man’s trash… Though consensus sequences are required for many valuable analyses, researchers in ancient DNA and microbial genomics in general must also acknowledge that single consensus sequences called from NGS data rarely reflect biological reality, even in clonal organisms. If the aDNA community is able to methodologically and culturally step outside the phylogenetics box, it could open the door to important, time structured evolutionary insight we lose by not considering the true variation of target organisms within our samples. The problem is, what is true variation in an ancient metagenomic sample?

Icebreaker mini-abstract 2 Building consensus alleles with awareness of position-specific damage probability To reconstruct the whole genome sequence of ancient bacteria, many recent studies use alignment-based approach by 1) aligning sequencing reads against a single reference sequence and 2) extracting consensus base from aligned reads. However, this strategy is artefact-prone when consensus site is decided by bases which are positioned on the ends of most aligned ancient reads where post-mortem damages are frequently observed. Therefore, we introduce a possible computational solution which could ameliorate the quality of reconstructed ancient bacterial genomes by precisely removing nucleotide bases of ancient reads which are possibly affected by post-mortem damage.

Session 3

Title Sorting and Recyling the trash: Defining authentication guidelines for the research community

Chair Clio Der Sarkissian and James Fellows Yates

Abstract Authenticating ancient metagenomic datasets has long been considered equally a nearly impossible and crucial task to demonstrate that ‘true’ ancient metagenomes can be characterised. Post-mortem fragmentation (typically via depurination) and damage base modification (via cytosine deamination) make aDNA analysis very sensitive to exogenous ‘contaminating’ DNA that can confound downstream analysis. Validation of results is therefore imperative not only to ensure scientific rigour, trust in the methods by researchers from related fields (e.g., anthropologists, archaeologists), but can also have important ethical consequences. Defining a consensus-based legitimate set of minimum-authentication criteria for ancient metagenomics is therefore crucial for gaining this trust in the field and facilitating further exploration of this type of data. Secondly, following standard guidelines by researchers can only be expected if these standards and guidelines follow FAIR principles: Finable, Accessible, Interoperable, and Reusable. Making sure authentication methods and reporting guidelines can be used and applied by any type of lab (big or small, specialist or generalist), can only come by making sure software and data is accessible, usable and most importantly understandable. In part one of this two-part session we will: i) get an overview of the obstacles and possible solutions of current and potential future approaches for authenticating metagenomic datasets, ii) begin to define a minimal authentication line of evidence and iii) discuss how to communicate this to collaborators inside and outside the ancient metagenomic community. In part two we will discuss i) how to improve accessibility and usability of ancient metagenomic datasets and software such as through collaboratively-generated benchmarking/comparative datasets as well as metadata reporting, and ii) how to ensure responsible and ethical research conduct.

Icebreaker presenters Sterling Wright, Nikolay Oskolkov, Nicolás Rascovan, James Fellows Yates, and Miriam Bravo López

Icebreaker 1 & 3 mini-abstract Improving current and future practices in ancient microbiome analyses In this presentation, we will discuss the different challenges in the bioinformatic processing and analysis of ancient microbiome data. This includes: i) compiling reference datasets and comparative analyses, ii) using standardized authentication tools, iii) identifying the limitations of current methods and analysis strategies, iv) recognizing community needs and designing more efficient methods, v) publishing studies that benchmark and validate strategies, vi) establishing database creation and handling, and vii) promoting open source and crowdsourcing initiatives. In particular, we will review how all of these topics affect authentication tools implemented in ancient microbiome studies, such as the commonly used SourceTracker. We hope that an open discussion on these issues will serve as a guide towards towards more robust and reproducible analysis in future research

Icebreaker 2 mini-abstract Improving authentication tools of ancient microbiomes In this presentation, I will discuss alternative ways of performing authentication analysis in ancient metagenomics data. Specifically, with low amounts of aDNA and a lack of good quality refernce genome, there is a need for going beyond traditional alignment and deamination pattern approaches, when addressing authenticity of each individual sequence. I will discuss how Machine and Deep Learning algorithms could help to infer authenticity of individual ancient metagenomics sequences in a reference genome free manner.

Icebreaker 4 mini-abstract: Towards standards in metadata reporting in ancient metagenomic studies: experiences from AncientMetagenomeDir One of the first and last part of most ancient metagenomic studies is to gathering or reporting basic metadata about samples and their sequences. A recent SPAAM community project established AncientMetagenomeDir, a comprehensive but lightwieght collection of lists of all published ancient metagenomic and microbial genome samples. In this icebreaker I will cover the challenges encountered during this data scraping exercise, and introduce one possible proposal to improving metadata reporting in our field.

Icebreaker 5 mini-abstract Ethics in ancient pathogen genomics The emergence of biological and historical interpretations through ancient DNA research, in particular through ancient pathogen genomics, can have important social, legal, and political consequences for individuals and communities. Therefore, the discussion of the best ethical practices within the field will allow us to promote responsible research on ancient pathogen genomics.

Session 5

Title Reuse the Refuse: Applying new analytical methods beyond current pratises

Chair Anna Fotakis and Alex Hübner

Abstract Throughout its history the field of ancient genomics has traditionally been playing catch-up by creatively adapting the advances achieved in modern genomics to their own types of datasets. Although this process has not been without setbacks, it steadily pushed the field forward and allowed it to apply the same methods as when using modern DNA. The newest technological advancement that took over metagenomics is de-novo assembly, whereby short reads from an environmental sample are assembled into novel microbial genomes. Its application to sequencing data from present-day samples has led to a rush of thousands of new metagenomic-assembled genomes (MAGs) from uncultivated microorganisms and allowed scientists to investigate the “unknown” at large scales for the first time. However, its application to sequencing data from ancient samples has just started and still requires thorough evaluation and fine-tuning of its parameters to make it suitable for this special type of sequencing data. In this session, we want to discuss the challenges that come along with applying de-novo assembly strategies to ancient metagenomic data and highlight potential pitfalls in the current workflows. Furthermore, we want to develop some guidelines on how to interpret this new type of results and establish which conclusions can be drawn from them, in order to avoid long-lasting mistakes that come along when adapting new methods to ancient genomics. Finally we would like to encourage novel ways of approaching datasets building upon previous sessions of SPAAM2 regarding public data sharing deposition.

Icebreaker presenters Maxime Borry and Antonio Fernandez-Guerra

Icebreaker mini-abstract 6: New take on ancient DNA metagenomics assembly validation
Given enough sequencing depth, it is now possible to assemble ancient DNA data into contigs using de novo assemblers. However, sorting the contigs originating from modern contamination, from contigs assembled from truly ancient DNA, can require a lot of manual work that does not scale easily with the immense number of assembled contigs. With the Pydamage approach of damage estimation and statistical testing, we show how to rapidly filter for contigs showing ancient DNA damage, to be used for further analysis.

Icebreaker mini-abstract 7 The adoption of modern metagenomic methods to explore ancient metagenomes and gain insight into their functional potential are starting to increase in their popularity amongst the community. However, we should act with caution and be aware of their limitations. I will go over some of the challenges of applying those approaches when analyzing ancient metagenomes and provide some ideas for developing computational methods tailored to our needs.

Session 6 - Future Steps for SPAAM

Abstract Now that we have identified common obstacles and discussed possible solutions for our field, what’s next? In this session we will summarise everything talked about in the last two days. We can identify potential common projects and ‘working commitees’ to begin work on these topics. Finally we will gather opinions on future meet ups, e.g. a SPAAM3.