Making mathematical online resources FAIR: at the example of small phylogenetic trees

Tabea Bacher, Marina Garrote-López, Christiane Görgen, Marius Neubert · 2025

Read Making mathematical online resources FAIR: at the example of small phylogenetic trees by Tabea Bacher, Marina Garrote-López, Christiane Görgen, Marius Neubert (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-10823.txt.

Introduction

Tabea Bacher, Marina Garrote-López, Christiane Görgen, and Marius Neubert reported on making a math data resource FAIR in 2025. FAIR data are findable, accessible, interoperable, and reusable. Their test case was an early web library of small phylogenetic trees, built in the early two thousands.

The team followed a three-part plan. They wrote software that lets users reproduce the database results. They built a friendlier web site with links to theory papers, code pieces, and stored outputs. All through the work, they documented each step to draw lessons for like projects. Their paper covers that third part, while the software and the site live at public addresses given in the text.

Catalog record

CREATORTabea Bacher, Marina Garrote-López, Christiane Görgen, Marius Neubert
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE31 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-10823

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Making mathematical online resources FAIR: at the example of small phylogenetic trees Tabea Bacher, Marina Garrote-López, Christiane Görgen, Marius Neubert Abstract We report on the process of taking an early 2000’s mathematical library, the Small Phylogenetic Trees, and transforming it into a FAIR, modern, and sustainable repository for data from algebraic phylogenetics. This process is based on a three-fold strategy: (1) writing a software package which enables the user to reproduce results of the database; (2) setting up a user-friendly new website with cross links to theoretical publications, code snippets, and serialized output of computations; and (3) all-the-while documenting the steps we take in order to derive lessons learned which may be generalised to other such projects. This paper addresses (3). (1) is found in [ 7 ] , and (2) is located at algebraicphylogenetics.org . Introduction The mathematical library Small Phylogenetic Trees [ 1 ] was first compiled in the early 2000s. It contains tables of algebraic invariants that characterize geometric properties of phylogenetic trees up to five taxa under group-based models. These tables are supplemented with code in Maple and Singular , with text files of computational output, and with explanations of terminology and notation. Figure 1 shows some screenshots. The library is an actively used online resource for researchers in algebraic phylogenetics, hosted at the main maintainer’s current academic institution [ 2 ] . In those early 2000s, the internet was being born. The now famous On-Line Encyclopedia of Integer Sequences oeis.org (OEIS) had been operational for about a decade, after having been a book for about two decades, and the L-functions and modular forms database lmfdb.org (LMFDB) was but a vague idea. The arXiv had yet to pass the half-million-article milestone, and professional mathematicians like Gowers and Tao were about to take to blogging. The initial release of the version-control system git had just been launched. Small Phylogenetic Trees’ aim to “make available in a unified [digital] format” known results about small tree models in algebraic phylogenetics was modern and visionary. The two decades that have passed since have seen a steep rise in the importance of digital resources for mathematicians and in the importance and potential of web technology. For instance, digital object identifiers (DOIs), introduced at the turn of the century, have all but taken over the role of ISBNs identifying books. GitHub and institutional gitlab instances have become standard in coding. It is now possible to display a large amount of data on a single page rather than on separate subpages. Open-source software for symbolic computations is available to match corporate technologies. Mathematical results can be linked to formalizations and papers. L a T e X can be displayed easily. And the FAIR Principles of findability, accessibility, interoperability, and reusability of digital research data [ 8 ] are on their rise throughout the sciences. These principles do not demand we use the newest and most shiny web tool, replacing old standards by ever new ones—much more cleverly, they demand we decide on a future-compatible way of treating our digital mathematical results in order to preserve their value and usability. Figure 1: These are screenshots of Small Phylogenetic Trees of January 2025. The first shows the landing page, the second an overview of all trees considered, and the third (after having navigated through two more subpages) data on the three-taxa phylogenetic tree with Jukes-Cantor model. By nowadays standards, the library Small Phylogenetic Trees is all but modern. Is it FAIR? Yes and no, as we will see in the next section. In this text, we take Small Phylogenetic Trees as an example of a valuable and influential mathematical resource with true and beautiful mathematical results but with a deprecated presentation, and we discuss how to use modern web technology to make it as FAIR as possible: taking the library way into the 21st century, while ensuring reproducibility, sustainability, and reusability of results. To do this, we pursue a three-fold strategy. For one, we launch a modern version of the original website, available now at algebraicphylogenetics.org . Two, we decide on one software that can provide computational proof for every result displayed on the website, and we use it to write a comprehensive package for algebraic phylogenetics; now available as AlgebraicPhylogenetics.jl [ 7 ] . And, three, we supplement these two work packages by the text you are reading, documenting our thought processes and their implementation. This latter point had been neglected in the original book chapter on Small Phylogenetic Trees but we believe it to be central. Our detailed project report, informed by the expertise of academic research-data managers and working mathematicians, can function as a solid bridge for many a no-longer-actively-maintained math website into a future where we raise and preserve all sorts of digital mathematical treasures. To achieve this, we let ourselves be guided by the FAIR Principles and give general recommendations on how to FAIR your own mathematical database. Changing the status quo Online mathematical libraries like OEIS, LMFDB or Small Phylogenetic Trees have now taken the role of twentieth-century look-up tables and classifications in book form. They may be used to test out and contradict conjectures, to discover for what proportions of well-studied cases a statement is true or false, or to make available results that are hard to obtain—thus saving time, brain, and energy resources. In addition, online resources are much more easily updated than their print counterparts, gathering information in a single location rather than spread out over a series of articles. For instance, computing the maximum likelihood degree of a three-taxa phylogenetic tree under the Jukes-Cantor model (which equals 23, as per the table in Figure 1 ) might be easily doable on a laptop. However, it involves Gröbner basis computations and these are notoriously hard even in relatively small cases. Already the maximum likelihood degree of the five-star tree under the Jukes-Cantor model is unknown. In fact, without redoing any computations on your own, Small Phylogenetic Trees tells you its variety is of dimension 6, degree 115, and lives in a 51-th dimensional linear space. The cardinality of the smallest set of generators defining its ideal is 175 and the generators have degree 4 and 5. These invariants reflect model complexity. Even more straightforwardly, if a researcher had a conjectured formula specifying the maximum likelihood degree of phylogenetic tree models, they could immediately test it with all such numbers available in the library. And they would know which, as of yet unknown, cases they could tackle to gather further evidence before attempting a theoretical proof. At the outset A lot of careful thought and effort has gone into Small Phylogenetic Trees, ensuring usability to this date. For instance, with the title as stated in the original publication [ 1 ] , the library can be found by any online search engine. Researchers may use the URL, provided by a trustworthy academic institution, to refer to Small Phylogenetic Trees or cite the corresponding book chapter. All original author names are listed on the landing page, central maintainers in bold at the top, and linked to their personal webpages: cf. Figure 1 . As in a catalog and as was typical in the early 2000s, the landing page

Read the complete text