Update requests queue

requests.jsonl

@@ -16,4 +16,4 @@
 {"id": "71e89fcc-43da-4b5a-9787-1c1f5e2b53cf", "type": "model", "title": "LULA-1.1 \u2014 omtx/lula-1.1", "description": "LULA-1.1 is a lightweight, sequence-only protein-ligand binding scorer from Om Therapeutics. It takes a protein amino-acid sequence and ligand SMILES and returns a binding score. There is no structure input, docking, or folding step.", "submitted_at": "2026-08-19T07:19:52.986Z", "source": "huggingscience.co", "status": "approved", "entry_id": "omtx-lula-1-1", "slug": "omtx/lula-1.1", "org_id": "omtx", "entry_type": "drug discovery", "tags": ["biology", "genomics"], "name": "LULA-1.1", "reviewed_at": "2026-08-19T07:22:38.393030+00:00", "pr_url": "https://github.com/georgiachanning/huggingscience-dot-co/pull/35"}
 {"id": "5785e471-2ea3-4bfc-9aa7-aa18ca03e9df", "type": "model", "title": "RudolfV 2 \u2014 Aignostics/rudolfv-2", "description": "Aignostics recently announced RudolfV 2, the best-performing openly available pathology foundation model to date. The model is built on a ViT-g architecture with approximately 1.1 billion parameters, and includes a dedicated post-training stage that to our knowledge is a first for pathology foundation models. We further distilled RudolfV 2 into two lightweight versions, RudolfV 2-B (ViT-B, 86M parameters) and RudolfV 2-S (ViT-S, 22M parameters), which preserve much of its performance at a fraction of the compute.\n\nHugging Face link: https://huggingface.co/collections/Aignostics/rudolfv-2\n\nMore info: https://www.aignostics.com/blog/rudolfv-2-a-state-of-the-art-open-weight-pathology-foundation-model", "submitted_at": "2026-08-26T20:51:59.985Z", "source": "huggingscience.co", "status": "approved", "entry_id": "aignostics-rudolfv-2", "slug": "Aignostics/rudolfv-2", "org_id": "Aignostics", "entry_type": "Pathology Foundation Model", "tags": ["medicine", "biotechnology"], "name": "RudolfV 2", "reviewed_at": "2026-08-27T09:56:41.056304+00:00", "pr_url": "https://github.com/georgiachanning/huggingscience-dot-co/pull/37"}
 {"id": "89d72710-1630-49d9-ab9d-13b38ada93a3", "type": "dataset", "title": "ratschlab/TCGA_virtual_spatial_transcriptomics_atlas", "description": "This is a new population-scale virtual spatial transcriptomics atlas, generated with DeepSpot-M by computationally augmenting TCGA retrospective histopathology archives spanning 28,664 images from 10,865 patients across 32 cancer types.\n\nThis resource opens new opportunities for cancer research and AI-driven precision medicine by enabling population-level studies of tissue biology through spatially resolved molecular maps covering 19k+ genes from H&E images.", "submitted_at": "2026-08-27T16:41:01.207Z", "source": "huggingscience.co", "status": "approved", "entry_id": "ratschlab-tcga-virtual-spatial-transcriptomics-atlas", "slug": "ratschlab/TCGA_virtual_spatial_transcriptomics_atlas", "org_id": "ratschlab", "entry_type": "Biology, Spatial Transcriptomics, Oncology, Pathology, Cancer", "tags": ["biology"], "reviewed_at": "2026-08-28T08:26:29.897488+00:00", "pr_url": "https://github.com/georgiachanning/huggingscience-dot-co/pull/38"}
-{"id": "a8b16a2f-e2d4-483c-9e29-bb0df4bdd2dd", "type": "dataset", "title": "ratschlab/DeepSpotM", "description": "DeepSpot-M, a multimodal foundation model that reads the full transcriptome (the activity of every gene) from medical images.\n\nMost tumours are never profiled with spatial transcriptomics, but nearly all are imaged. DeepSpot-M is bridging that gap and as a first demonstration, we used it to build a virtual spatial transcriptomics atlas of 28,664 TCGA slides across 32 cancers, the largest resource of its kind.\n\nOlder methods predict a small, fixed set of genes. We ask a different question instead: \"how much is gene X expressed in this image?\" Every gene gets its own profile, built from its DNA and RNA sequence, its protein activity, how it behaves at the single cell level, and what's known about it in the literature. Because the model reasons over these profiles rather than memorizing a gene list, it can predict genes it never saw during training just as accurately as ones it did. \n\nTrained on one of the largest oncology datasets to date, DeepSpot-M generalizes to cancers it has never seen and even beats specialist models trained directly on those cancers. Given just a few slides from a new cohort, it adapts further and keeps improving.\n\nThe same system also enables:\n\ud83d\udd2c Extending targeted single-cell spatial assays toward full transcriptome-wide coverage\n\ud83e\ude7a Restoring degraded or low-quality spatial transcriptomics data\n\ud83e\uddea In silico variant effect mapping conditioned on morphology\n\ud83d\udcac Natural language querying of tissue biology", "submitted_at": "2026-08-27T16:42:40.107Z", "source": "huggingscience.co", "status": "pending", "entry_id": "ratschlab-deepspotm", "slug": "ratschlab/DeepSpotM", "org_id": "ratschlab", "entry_type": "Biology, Spatial Transcriptomics, Model, Oncology", "tags": ["biology"]}
+{"id": "a8b16a2f-e2d4-483c-9e29-bb0df4bdd2dd", "type": "dataset", "title": "ratschlab/DeepSpotM", "description": "DeepSpot-M, a multimodal foundation model that reads the full transcriptome (the activity of every gene) from medical images.\n\nMost tumours are never profiled with spatial transcriptomics, but nearly all are imaged. DeepSpot-M is bridging that gap and as a first demonstration, we used it to build a virtual spatial transcriptomics atlas of 28,664 TCGA slides across 32 cancers, the largest resource of its kind.\n\nOlder methods predict a small, fixed set of genes. We ask a different question instead: \"how much is gene X expressed in this image?\" Every gene gets its own profile, built from its DNA and RNA sequence, its protein activity, how it behaves at the single cell level, and what's known about it in the literature. Because the model reasons over these profiles rather than memorizing a gene list, it can predict genes it never saw during training just as accurately as ones it did. \n\nTrained on one of the largest oncology datasets to date, DeepSpot-M generalizes to cancers it has never seen and even beats specialist models trained directly on those cancers. Given just a few slides from a new cohort, it adapts further and keeps improving.\n\nThe same system also enables:\n\ud83d\udd2c Extending targeted single-cell spatial assays toward full transcriptome-wide coverage\n\ud83e\ude7a Restoring degraded or low-quality spatial transcriptomics data\n\ud83e\uddea In silico variant effect mapping conditioned on morphology\n\ud83d\udcac Natural language querying of tissue biology", "submitted_at": "2026-08-27T16:42:40.107Z", "source": "huggingscience.co", "status": "rejected", "entry_id": "ratschlab-deepspotm", "slug": "ratschlab/DeepSpotM", "org_id": "ratschlab", "entry_type": "Biology, Spatial Transcriptomics, Model, Oncology", "tags": ["biology"], "reviewed_at": "2026-08-28T09:06:20.092716+00:00", "reject_reason": "Added with the ratschlab dataset "}