Bilkent CompGen

Several changes in our group in the 2026-2027 Academic Year. Most recently, Zülal has finished her Ph.D., and new members have joined our group. Here is the @BilkentCompGen, October 2026.

Preprint at bioRxiv: biorxiv.org/content/10.64898… Code: github.com/Malfoy/Zindex #bioinformatics #genomics #datastructures

• All 2.76M genomes of AllTheBacteria indexed into 406 GB with 43.8 GB peak memory • Full RefSeq (564k assemblies) indexed in under 22 hours • On 268k M. tuberculosis assemblies: 5.5x smaller index, 9.5x faster build and 8x less memory than state-of-the-art at the same FPR

In this paper, we introduce Zindex, an approximate colored index built on interleaved ZOR filters, previously developed by Antoine, with a color-preserving merging scheme and support for minimizers, gapped k-mers, and randstrobes.

Huge shout-out to Antoine — this collaboration started over a chat about Tuna's poster at RECOMB 2026 in Thessaloniki, and it's been a pleasure to see it turn into this.

Excited to share our new preprint: "Toward space-optimal approximate colored indexes at scale" This is the work of Tuna Okçu (@TunaO123), who led the project together with Antoine Limasset (@BQPMalfoy). They are a powerful duo!

This work was completed within the @ProjectBioPIM, in collaboration with our partners at @SAFARI_ETH_CMU, namely Konstantina Koliogeorgi, @FirtinaC, and @_onurmutlu_ #organicpost

... memory usage, memory bus utilization, and execution speed. Through our rigorous analysis, we aim to provide insights for hardware acceleration of pangenomics workloads, such as domain-specific accelerators and processing-in-memory architectures.

Led by @zulal_b and contributed by Berkan Şahin, Klea Zambaku, and @rica01234, in this paper we systematically analyze modern de Bruijn graph-based tools for pangenomics and characterize their performance and resource footprint, including energy efficiency, ...

Our new paper, "De Bruijn graphs for pangenomics: in-depth performance benchmarking of de Bruijn graph-based tools for read mapping," is now published in Briefings in Bioinformatics. doi.org/10.1093/bib/bbag440

Another one of our alumni, Damla Senol Cali, visited us from San Diego. Thanks to Eleazar Eskin, Stacey Beggs, and Kevin Kim for organizing yet another great meeting. One more @UCLA_CGSI in the books, already looking forward to the next!

We also met with our alumnus Mohammed Alser, and three of his brilliant students, Adrian Tkachenko, Sepehr Salem, and Mohammed Nayeem Uddin. It was also great to reconnect with many of our friends and to sow the seeds of several collaborative projects.

Tuna Okçu presented his work on metagenomics using binary fuse filters, and Banu Cavlak presented her work on pre-basecalling filtering for targeted sequencing applications.

@calkan_cs gave two talks: a tutorial on accelerating bioinformatics workloads, and a research paper presentation on string processing using locally consistent parsing.

We attended the @UCLA_CGSI Computational Genomics Summer Institute 2026 hosted by UCLA.

... PG2 provides the algorithms and an intuitive web-based interface for the seamless and scalable visual analysis of pangenome graphs. Read the paper here: link.springer.com/article/10… And test it out at: pg2.cs.bilkent.edu.tr/

Led by Ugur Dogrusoz, and completed by a brilliant undergraduate student (at the time), Görkem Kadir Solun (now pursuing M.Sc. at ETH Zurich), and contributed to by my PhD student Zülal Bingöl ...

Our new paper, "PG2: algorithms and a web-based tool for effective layout and visual analysis of pangenome graphs," in collaboration with the Bilkent i-Vis Research Lab, is officially out in BMC Bioinformatics.

Additionally, Klea Zambaku presented a poster about our work within the BioPIM Project. We also met with our alumni, Ezgi Ebren, Can Firtina, and Mohammed Alser, as well as many friends around the world. Looking forward to the next RECOMB in Toronto!

A few weeks ago, we attended the RECOMB 2026 Conference and its associated Satellite Conferences RECOMB-Seq and RECOMB-Arch in Thessaloniki, Greece. Arda İçöz and Tuna Okçu presented their works in person, and Berkan Şahin presented his work remotely.