One person,
one long
paper trail.

Everything that shipped, dated the day it landed.

Read straight from the repository's own Changelog at build time, so this page cannot flatter the project by being out of date.

1. 13 Nov 202602

   • Genetic circuits can now predict the whole population of cells each input state produces, drawn as flow-cytometry histograms from the gate library's own measurements.

   • The genetic circuit panel now shows how a design was produced: each stage from specification to DNA with what it actually computed, beside the strain and growth conditions its parts were measured in.

2. 12 Nov 202602

   • The alignment viewer can draw a consensus histogram over the columns by UGENE's Default, Strict, Levitsky or ClustalW rule, and its overview strip shows and drags the visible window.

   • The alignment viewer colours residues by Jalview's named schemes (Clustal, Zappo, Taylor, hydrophobicity, percentage identity, BLOSUM62), sorts by pairwise identity to a chosen row, and shows row, column and residue number in the status…

3. 11 Nov 202601

   • Variant calls: a VCF 4.4 single breakend (such as .CCCCCCG) is now classed as a breakend and refused by Apply instead of being written into the sequence, and the Summary counts only the INFO and FORMAT fields the file declares.

4. 7 Nov 202601

   • Flow cytometry gains kinetics (a channel's median, mean, percentile or % above a threshold over acquisition time, with per-range peak, slope and area) and derived parameters (new channels from expressions such as a calcium ratio).

5. 6 Nov 202601

   • Flow cytometry: open an FCS file, compensate it with its own spillover, apply a Gating-ML 2.0 gating strategy and read every population's size and per-channel statistics, with a density plot of any two channels.

6. 5 Nov 202602

   • Ligation-independent cloning now designs its primers as well as predicting the reaction: LIC, SLIC, In-Fusion and CPEC each at their own authority's homology arm, and USER as the complementary deoxyuridine tail pair its method really…

   • The MCP server lists each tool more briefly (the full text is one describe_tools call away), offers a search-first endpoint at /mcp/discover, and returns very large answers in windows you can page through instead of all at once.

7. 26 Oct 202601

   • Bench Calculators now carry the vendor's own PCR setup for every catalogued polymerase (NEB, Takara Bio, QIAGEN), work out primer and magnesium volumes from your stock, and gained Serial dilution and Mass ↔ moles tabs.

8. 25 Oct 202605

   • Cloning without a ligase now names the method: LIC, SLIC, In-Fusion, CPEC or USER, each run with the chemistry and the homology arm its own authority states.

   • Lab notebook: run any analysis inside an entry.

     A run cell computes on the open record, shows its result with units, says when it went stale, feeds later cells, draws on the map and travels into Word and PDF exports.

   • The alignment panel finds redundant rows at a chosen identity and removes them in one undoable step, and plots a PCA of the rows under percent identity, BLOSUM62, PAM250 or a DNA matrix; both match Jalview's own numbers.

   • The variant table now says what each call does to a gene: a Consequence column with the Sequence Ontology term, Ensembl's impact rating and HGVS notation, sorted most severe first and checked against SnpEff.

   • You can now edit an alignment by hand: insert and delete gaps in one row or every row, trim the ends, remove all-gap columns, a row's gaps or a row, with undo and redo, and residue ranges renumbered as you go.

9. 24 Oct 202612

   • Back up every project and your lab notebook to one file from the projects page, and restore it in any browser without overwriting newer work.

   • Clearing your browser's site data no longer hides projects that were saved in the browser's private storage: they are rebuilt on the next load.

   • Codon Optimization can lay a fragment out as IDT's gene-fragment plate upload and download it as a spreadsheet, refusing fragments outside the product's length range.

   • Tool panels can now save their settings as a named profile, apply one, and export or import it as a file to share with colleagues.

   • Export the RNA-seq gene table and every discovery panel's table as a file, save the domain scanner's motif and profile readings to the notebook, and see a protein's curated UniProt sites on the map.

   • FIG now draws a gel or table it produced right in its reply, and when it opens a document it just made, the next step works on that document.

   • Gibson Assembly can now open one closed vector at several enzyme sites and place an insert in every gap in a single step.

   • Project search now finds bases with mismatches, IUPAC codes or as amino acids in six frames, and can search only inside a named feature.

   • Protein Domains lists every curated UniProt modification, a primer tail can come from your saved oligos, the file manager compares the facts of several files in one table, and Reagent Audit, Trade-offs, Population Genetics and Power now…

   • The app's first visit downloads 63.5 MB in the background instead of 92.3 MB; nine large per-tool tables now download the first time their tool opens and stay available offline after that.

   • The rendered map can now be drawn linear, with the features in lanes under a ruler and, on request, both strands' bases for the whole record or just your selection; maps sent to an AI assistant use the editor's feature colours.

   • Translations can be coloured by Zappo residue groups or membrane tendency, and an expanded enzyme row in Properties draws its cut lines across the sequence.

10. 23 Oct 202610

    • An opened paper or PDF can list the UniProt and RefSeq accessions it names, each checked against its archive, and open any of them as a record in the project.

    • Every tool reached over the API or MCP now answers with the same reading its panel shows - the conclusion, what undermines it and how it was computed - or says in its contract why it has none.

    • Gene set enrichment can now run as fgsea itself runs it, multilevel or with plain permutations, so a seed gives fgsea's own p-values.

    • Gene set enrichment can now use clusterProfiler's plain permutation tests: random gene samples or relabelled rankings, a fixed number of draws or adaptive batches, each matching clusterProfiler's p-values for a seed.

    • Gene sets can now be fetched live for your organism: Gene Ontology terms from the GO Consortium or Reactome pathways for your genes, each stamped with its source, release and date.

    • RNA-seq: a new camera question tests whether each gene set moves more than the rest of the genes, with the correlation between its genes allowed for, from a voom fit of the counts or from a ranked statistic.

    • RNA-seq over-representation can weight each gene, for instance by the DE run's mean counts, and test with Wallenius' distribution so well-counted sets are not called just for being easy to detect.

    • Scanning a chromosomal microarray for findings no longer freezes the page: the results draw once, in about a quarter of the time.

    • The copilot now works out of the box again: with no key of your own it tries the free models in turn and keeps the first one that actually answers, instead of opening on one that refuses.

    • The lab notebook now takes figures from any plot or tab, writes a methods section from your builds, exports to Word and PDF, keeps a version history per experiment, and shows up in project search.

11. 22 Oct 202609

    • A paper found in Databases can now be cited into the notebook, or opened in full as a project file when it is open access, with its licence and source.

    • An article opened from Europe PMC now keeps its figure captions, and clicking a figure reference in the text goes to that figure.

    • Find words in a PDF, step through the pages that hold them, and open a project-search hit straight on its page.

    • Highlight passages in a PDF and add a note to each; highlights stay with the file in the project and are found again by their words.

    • RNA-seq differential expression now draws DESeq2's sample checks: a PCA of the variance-stabilized (or rlog) counts coloured by any sample variable, and the sample-to-sample distance heatmap with its clustering, every value matching R.

    • RNA-seq differential expression now offers DESeq2's glmGamPoi dispersions, the single-cell path, with glmGamPoi's quasi-likelihood F test under the likelihood-ratio design; every gene matches DESeq2 and glmGamPoi in R.

    • RNA-seq differential expression takes tximport's average transcript lengths and normalizes by gene and sample, as DESeq2 does.

    • Select words in a PDF or an opened open-access article and quote them into the notebook with the paper's citation and page.

    • The RNA-seq panel can fit DESeq2's local dispersion trend, and takes observation weights (such as zinbwave's for single cells) with t-distributed Wald tests; a single-cell starter opens on them.