Thrilled to announce that my first 1st-author paper has been published in Briefings in Bioinformatics! 🎉🧬 @OUPAcademic ​Introducing QTFPred: A robust quantum-classical hybrid model that predicts transcription factor binding at base resolution. 🧬⚛️ ​📄 Paper:

工学的研究が素晴らしいのは、始めた瞬間に論文化を確定できること。特に論文を書く時間は必要なく論文化できることですね。あくまで私流。結果を得てから最短１日で論文は完成します。IF4くらいは普通に通ります。

「ゲノム言語モデルの現在と未来」を語りましょう! 第48回日本分子生物学会年会、12月4日(木) 19:15-20:30、パシフィコ横浜 会議センター 3F (第8会場)のフォーラムにパネルディスカッションに登壇します。#MBSJ2025

@OxUniPress @PennyLaneAI @kosuke_mitarai #QTFPred integrates a Parameterized Quantum Circuit (PQC) directly into a DL pipeline. ​The Mechanism: 1️⃣ Variational Params : Trainable rotation angles embedded in the circuit. 2️⃣ Exponential Feature Space: By exploiting superposition, the circuit processes features in a massive

弊社同期のうち東工大、阪大卒で4人ほど非常に頭が切れる優秀な奴がいたが、上司ガチャ外れたのか久々に会話したらこの内3人はすっかり頭のキレが無くなっていた 特に内1人は病んで1年強休職した後で、普通の会話のキャッチボールすら怪しくなっていた

【遂に出た】Briefings in Bioinformatics誌から筆頭論文Published！🧬 "QTFPred"を世に放ちました🧬🔯 量子機械学習によって転写因子-DNA結合を高精度予測します 量子計算、量子情報の可能性をゲノムインフォマティクスの分野で開拓した僕のデビュー作です Check it out! https://t.co/IRGkxhU0iu

【注目プレスリリース】BlueMemeと九州大学、量子AIを活用した先進的ゲノム解析技術の研究成果が国際学術誌に掲載 創薬・医療研究の新たな可能性を切り拓く量子機械学習モデル「QTFPred」を発表 / 九州大学

BlueMemeと九州大学、量子AIを活用した先進的ゲノム解析技術の研究成果が国際学術誌に掲載 創薬・医療研究の新たな可能性を切り拓く量子機械学習モデル「QTFPred」を発表 📗 https://t.co/HxU7QR1rV0 #九州大学 #九大研究成果 #九大 #KyushuU https://t.co/ZgTPOlQFjC

@OxUniPress important points about implementation: quantum conv layer is written by pennylane @PennyLaneAI model is based on Quantum Circuit Learning (QCL) @kosuke_mitarai

This work opens the door for Quantum ML in genomics, especially for rare targets🧬. ​Huge thanks to my supervisors and co-authors for their guidance on phd journey! 😆 ​📖 Read here: https://t.co/XtzQnCR5SY 💻 Code: https://t.co/9x0H2bZyeQ ​@OxUniPress

Benchmarked on 49 datasets, QTFPred achieved state-of-the-art accuracy in: ✅ 92% of binary classification tasks ✅ 96% of signal prediction tasks ​ Crucially, it significantly outperforms baselines (BPNet, FCNsignal) specifically in data-sparse groups (<10k peaks).

New: Quantum Computation ⚛️ ​ We replaced the standard first layer with a Quantum Convolutional (QConv) layer. ​By mapping DNA sequences🧬 into an exponentially large quantum feature space, QTFPred captures complex patterns more efficiently than classical kernels.

We needed a new approach🧬⚛️. Deep learning models hunger for data. But in genomics, many TFs have sparse ChIP-seq peaks. In fact, ~45% of ENCODE experiments have <10k peaks. ​Conventional models often overfit or fail to capture motifs in these "low-data" scenarios.

#ASHG2025 #HumanGenetics #Genomics #MachineLearning #QuantumComputing #Bioinformatics #KyushuUniversity #Research #PosterPresentation #Japan

This work is part of my doctoral research at Kyushu University (Nagasaki Lab). We anticipate this paper will be published soon, so if you are interested in this research, please follow me for future updates!🧬🧬🧬

I successfully completed my poster presentation at the ASHG 2025 Annual Meeting in Boston! ​I shared our findings on Quantum machine learning for robust transcription factor binding prediction. Thank you to everyone who visited my poster and engaged in stimulating discussions.