<HashMap><database>biostudies-literature</database><scores/><additional><submitter>Gasparotto P</submitter><funding>Swiss Data Science Center</funding><pagination>931-944</pagination><full_dataset_link>https://www.ebi.ac.uk/biostudies/studies/S-EPMC11299607</full_dataset_link><repository>biostudies-literature</repository><omics_type>Unknown</omics_type><volume>57(Pt 4)</volume><pubmed_abstract>Serial crystallography (SX) involves combining observations from a very large number of diffraction patterns coming from crystals in random orientations. To compile a complete data set, these patterns must be indexed (&lt;i>i.e.&lt;/i> their orientation determined), integrated and merged. Introduced here is &lt;i>TORO&lt;/i> (&lt;i>Torch&lt;/i>-powered robust optimization) &lt;i>Indexer&lt;/i>, a robust and adaptable indexing algorithm developed using the &lt;i>PyTorch&lt;/i> framework. &lt;i>TORO&lt;/i> is capable of operating on graphics processing units (GPUs), central processing units (CPUs) and other hardware accelerators supported by &lt;i>PyTorch&lt;/i>, ensuring compatibility with a wide variety of computational setups. In tests, &lt;i>TORO&lt;/i> outpaces existing solutions, indexing thousands of frames per second when running </pubmed_abstract><journal>Journal of applied crystallography</journal><pubmed_title>&lt;i>TORO Indexer&lt;/i>: a &lt;i>PyTorch&lt;/i>-based indexing algorithm for kilohertz serial crystallography.</pubmed_title><pmcid>PMC11299607</pmcid><funding_grant_id>C19-03 (RED-ML project)</funding_grant_id><pubmed_authors>Assmann G</pubmed_authors><pubmed_authors>Bejar B</pubmed_authors><pubmed_authors>Stadler HC</pubmed_authors><pubmed_authors>Ashton AW</pubmed_authors><pubmed_authors>Leonarski F</pubmed_authors><pubmed_authors>Barba L</pubmed_authors><pubmed_authors>Gasparotto P</pubmed_authors><pubmed_authors>Mendonca H</pubmed_authors><pubmed_authors>Janousch M</pubmed_authors></additional><is_claimable>false</is_claimable><name>&lt;i>TORO Indexer&lt;/i>: a &lt;i>PyTorch&lt;/i>-based indexing algorithm for kilohertz serial crystallography.</name><description>Serial crystallography (SX) involves combining observations from a very large number of diffraction patterns coming from crystals in random orientations. To compile a complete data set, these patterns must be indexed (&lt;i>i.e.&lt;/i> their orientation determined), integrated and merged. Introduced here is &lt;i>TORO&lt;/i> (&lt;i>Torch&lt;/i>-powered robust optimization) &lt;i>Indexer&lt;/i>, a robust and adaptable indexing algorithm developed using the &lt;i>PyTorch&lt;/i> framework. &lt;i>TORO&lt;/i> is capable of operating on graphics processing units (GPUs), central processing units (CPUs) and other hardware accelerators supported by &lt;i>PyTorch&lt;/i>, ensuring compatibility with a wide variety of computational setups. In tests, &lt;i>TORO&lt;/i> outpaces existing solutions, indexing thousands of frames per second when running </description><dates><release>2024-01-01T00:00:00Z</release><publication>2024 Aug</publication><modification>2025-04-26T05:26:21.212Z</modification><creation>2025-04-06T11:33:04.242Z</creation></dates><accession>S-EPMC11299607</accession><cross_references><pubmed>39108821</pubmed><doi>10.1107/S1600576724003182</doi></cross_references></HashMap>