Release Notes¶

Note

Also see the release notes of anndata.

Version 1.4¶

1.4.6 2020-03-17¶

Functionality in external

sam() self-assembling manifolds [Tarashansky19] PR 903 A Tarashansky
harmony_timeseries() for trajectory inference on discrete time points PR 994 A Mousa
wishbone() for trajectory inference (bifurcations) PR 1063 A Mousa

Code design

violin now reads .uns['colors_...'] PR 1029 michalk8

Bug fixes

adapt ingest() for UMAP 0.4 PR 1038 PR 1106 S Rybakov
compat with matplotlib 3.1 and 3.2 PR 1090 I Virshup, P Angerer
fix PAGA for new igraph PR 1037 P Angerer
fix rapids compat of louvain PR 1079 LouisFaure

1.4.5 2019-12-30¶

Please install scanpy==1.4.5.post3 instead of scanpy==1.4.5.

New functionality

ingest() maps labels and embeddings of reference data to new data → tutorial: integrating-data-using-ingest PR 651 S Rybakov, A Wolf
queries recieved many updates including enrichment through gprofiler and more advanced biomart queries PR 467 I Virshup
set_figure_params() allows setting figsize and accepts facecolor='white', useful for working in dark mode A Wolf

Code design

downsample_counts now always preserves the dtype of it’s input, instead of converting floats to ints PR 865 I Virshup
allow specifying a base for log1p() PR 931 G Eraslan
run neighbors on a GPU using rapids PR 850 T White
param docs from typed params P Angerer
embedding_density() now only takes one positional argument; similar for embedding_density(), which gains a param groupby PR 965 A Wolf
webpage overhaul, ecosystem page, release notes, tutorials overhaul PR 960 PR 966 A Wolf

Warning

changed default solver in pca() from auto to arpack
changed default use_raw in score_genes() from False to None

1.4.4 2019-07-20¶

New functionality

scanpy.get adds helper functions for extracting data in convenient formats PR 619 I Virshup

Bug fixes

Stopped deprecations warnings from AnnData 0.6.22 I Virshup

Code design

normalize_total() gains param exclude_highly_expressed, and fraction is renamed to max_fraction with better docs A Wolf

1.4.3 2019-05-14¶

Bug fixes

neighbors() correctly infers n_neighbors again from params, which was temporarily broken in v1.4.2 I Virshup

Code design

calculate_qc_metrics() is single threaded by default for datasets under 300,000 cells – allowing cached compilation PR 615 I Virshup

1.4.2 2019-05-06¶

New functionality

combat() supports additional covariates which may include adjustment variables or biological condition PR 618 G Eraslan
highly_variable_genes() has a batch_key option which performs HVG selection in each batch separately to avoid selecting genes that vary strongly across batches PR 622 G Eraslan

Bug fixes

rank_genes_groups() t-test implementation doesn’t return NaN when variance is 0, also changed to scipy’s implementation PR 621 I Virshup
umap() with init_pos='paga' detects correct dtype A Wolf
louvain() and leiden() auto-generate key_added=louvain_R upon passing restrict_to, which was temporarily changed in 1.4.1 A Wolf

Code design

neighbors() and umap() got rid of UMAP legacy code and introduced UMAP as a dependency PR 576 S Rybakov

1.4.1 2019-04-26¶

New functionality

Scanpy has a command line interface again. Invoking it with scanpy somecommand [args] calls scanpy-somecommand [args], except for builtin commands (currently scanpy settings) PR 604 P Angerer
ebi_expression_atlas() allows convenient download of EBI expression atlas I Virshup
marker_gene_overlap() computes overlaps of marker genes M Luecken
filter_rank_genes_groups() filters out genes based on fold change and fraction of cells expressing genes F Ramirez
normalize_total() replaces normalize_per_cell(), is more efficient and provides a parameter to only normalize using a fraction of expressed genes S Rybakov
downsample_counts() has been sped up, changed default value of replace parameter to False PR 474 I Virshup
embedding_density() computes densities on embeddings PR 543 M Luecken
palantir() interfaces Palantir [Setty18] PR 493 A Mousa

Code design

.layers support of scatter plots F Ramirez
fix double-logarithmization in compute of log fold change in rank_genes_groups() A Muñoz-Rojas
fix return sections of docs P Angerer

Version 1.3¶

1.3.8 2019-02-05¶

read_10x_h5() throws more stringent errors and doesn’t require speciying default genomes anymore. PR 442 and PR 444 I Vishrup

1.3.7 2019-01-02¶

Major updates

one can import scanpy as sc instead of import scanpy.api as sc, see scanpy

New functionality

combat() reimplements Combat for batch effect correction [Johnson07] [Leek12], heavily based on the Python implementation of [Pedersen12], but with performance improvements PR 398 :smaller:`M Lange
phenograph() wraps the graph clustering package Phenograph [Levine15] A Mousa

1.3.6 2018-12-11¶

Major updates

a new plotting gallery for Visualizing marker genes F Ramirez
tutorials are integrated on ReadTheDocs, Preprocessing and clustering 3k PBMCs and Trajectory inference for hematopoiesis in mouse

Interactive exploration of analysis results through manifold viewers

CZI’s cellxgene directly reads .h5ad files the cellxgene developers
the UCSC Single Cell Browser requires exporting via cellbrowser() M Haeussler

Code design

highly_variable_genes() supersedes filter_genes_dispersion(), it gives the same results but, by default, expects logarithmized data and doesn’t subset

1.3.5 2018-12-09¶

uncountable figure improvements PR 369 F Ramirez

1.3.4 2018-11-24¶

leiden() wraps the recent graph clustering package by [Traag18] K Polanski
bbknn() wraps the recent batch correction package [Polanski19] K Polanski
calculate_qc_metrics() caculates a number of quality control metrics, similar to calculateQCMetrics from Scater [McCarthy17] I Virshup

1.3.3 2018-11-05¶

Major updates

a fully distributed preprocessing backend T White and the Laserson Lab

Code design

read_10x_h5() and read_10x_mtx() read Cell Ranger 3.0 outputs PR 334 Q Gong

Note

Also see changes in anndata 0.6.

changed default compression to None in write_h5ad() to speed up read and write, disk space use is usually less critical
performance gains in write_h5ad() due to better handling of strings and categories S Rybakov

1.3.1 2018-09-03¶

RNA velocity in single cells [Manno18]

Scanpy and AnnData support loom’s layers so that computations for single-cell RNA velocity [Manno18] become feasible S Rybakov and V Bergen
the package scvelo perfectly harmonizes with Scanpy and is able to process loom files with splicing information produced by Velocyto [Manno18], it runs a lot faster than the count matrix analysis of Velocyto and provides several conceptual developments (preprint to come)

Plotting (Generic)

dotplot() for visualizing genes across conditions and clusters, see here PR 199 F Ramirez
heatmap() for pretty heatmaps PR 175 F Ramirez
violin() produces very compact overview figures with many panels PR 175 F Ramirez

There now is a section on imputation in external:

magic() for imputation using data diffusion [vanDijk18] PR 187 S Gigante
dca() for imputation and latent space construction using an autoencoder [Eraslan18] PR 186 G Eraslan

Version 1.2¶

1.2.1 2018-06-08¶

Plotting of Generic marker genes and quality control.

highest_expr_genes() for quality control; plot genes with highest mean fraction of cells, similar to plotQC of Scater [McCarthy17] PR 169 F Ramirez

1.2.0 2018-06-08¶

paga() improved, see PAGA; the default model changed, restore the previous default model by passing model='v1.0'

Version 1.1¶

1.1.0 2018-06-01¶

set_figure_params() by default passes vector_friendly=True and allows you to produce reasonablly sized pdfs by rasterizing large scatter plots
draw_graph() defaults to the ForceAtlas2 layout [Jacomy14] [Chippada18], which is often more visually appealing and whose computation is much faster S Wollock
scatter() also plots along variables axis MD Luecken
pca() and log1p() support chunk processing S Rybakov
regress_out() is back to multiprocessing F Ramirez
read() reads compressed text files G Eraslan
mitochondrial_genes() for querying mito genes FG Brundu
mnn_correct() for batch correction [Haghverdi18] [Kang18]
phate() for low-dimensional embedding [Moon17] S Gigante
sandbag(), cyclone() for scoring genes [Scialdone15] [Fechtner18]

Version 1.0¶

1.0.0 2018-03-30¶

Major updates

Scanpy is much faster and more memory efficient: preprocess, cluster and visualize 1.3M cells in 6h, 130K cells in 14min, and 68K cells in 3min
the API gained a preprocessing function neighbors() and a class Neighbors() to which all basic graph computations are delegated

Warning

Upgrading to 1.0 isn’t fully backwards compatible in the following changes

the graph-based tools louvain() dpt() draw_graph() umap() diffmap() paga() require prior computation of the graph: sc.pp.neighbors(adata, n_neighbors=5); sc.tl.louvain(adata) instead of previously sc.tl.louvain(adata, n_neighbors=5)
install numba via conda install numba, which replaces cython
the default connectivity measure (dpt will look different using default settings) changed. setting method='gauss' in sc.pp.neighbors uses gauss kernel connectivities and reproduces the previous behavior, see, for instance in the example paul15.
namings of returned annotation have changed for less bloated AnnData objects, which means that some of the unstructured annotation of old AnnData files is not recognized anymore
replace occurances of group_by with groupby (consistency with pandas)
it is worth checking out the notebook examples to see changes, e.g. the seurat example.
upgrading scikit-learn from 0.18 to 0.19 changed the implementation of PCA, some results might therefore look slightly different

Further updates

UMAP [McInnes18] can serve as a first visualization of the data just as tSNE, in contrast to tSNE, UMAP directly embeds the single-cell graph and is faster; UMAP is also used for measuring connectivities and computing neighbors, see neighbors()
graph abstraction: AGA is renamed to PAGA: paga(); now, it only measures connectivities between partitions of the single-cell graph, pseudotime and clustering need to be computed separately via louvain() and dpt(), the connectivity measure has been improved
logistic regression for finding marker genes rank_genes_groups() with parameter method='logreg'
louvain() provides a better implementation for reclustering via restrict_to
scanpy no longer modifies rcParams upon import, call settings.set_figure_params to set the ‘scanpy style’
default cache directory is ./cache/, set settings.cachedir to change this; nested directories in this are avoided
show edges in scatter plots based on graph visualization draw_graph() and umap() by passing edges=True
downsample_counts() for downsampling counts MD Luecken
default 'louvain_groups' are called 'louvain'
'X_diffmap' contains the zero component, plotting remains unchanged

Version 0.4¶

0.4.4 2018-02-26¶

embed cells using umap() [McInnes18] PR 92 G Eraslan
score sets of genes, e.g. for cell cycle, using score_genes() [Satija15] D Cittaro

0.4.3 2018-02-09¶

clustermap(): heatmap from hierarchical clustering, based on seaborn.clustermap() [Waskom16]
only return matplotlib.axes.Axes in plotting functions of sc.pl when show=False, otherwise None

0.4.2 2018-01-07¶

amendments in PAGA and its plotting functions

0.4.0 2017-12-23¶

export to SPRING [Weinreb17] for interactive visualization of data: spring tutorial

Version 0.3¶

0.3.2 2017-11-29¶

finding marker genes via rank_genes_groups_violin() improved, see issue 51

0.3.0 2017-11-16¶

AnnData gains method concatenate()
AnnData is available as the separate anndata package
results of PAGA simplified

Version 0.2¶

0.2.9 2017-10-25¶

Initial release of the new trajectory inference method PAGA

paga() computes an abstracted, coarse-grained (PAGA) graph of the neighborhood graph
paga_compare() plot this graph next an embedding
paga_path() plots a heatmap through a node sequence in the PAGA graph

0.2.1 2017-07-24¶

Scanpy includes preprocessing, visualization, clustering, pseudotime and trajectory inference, differential expression testing and simulation of gene regulatory networks. The implementation efficiently deals with datasets of more than one million cells.

Version 0.1¶

0.1.0 2017-05-17¶

Scanpy computationally outperforms and allows reproducing both the Cell Ranger R kit’s and most of Seurat’s clustering workflows.