update references

README.md

@@ -8,7 +8,7 @@
 [![Downloads](https://static.pepy.tech/badge/liana)](https://pepy.tech/project/liana)
 <!-- badges: end -->
 
-LIANA+ is a scalable framework that integrates and extends existing methods and knowledge to study cell-cell communication in single-cell, spatially-resolved, and multi-modal omics data. It is part of the [scverse ecosystem](https://github.com/scverse), and relies on [AnnData](https://github.com/scverse/anndata) & [MuData](https://github.com/scverse/mudata) objects as input.
+LIANA+ is a scalable framework that adapts and extends existing methods and knowledge to study cell-cell communication in single-cell, spatially-resolved, and multi-modal omics data. It is part of the [scverse ecosystem](https://github.com/scverse), and relies on [AnnData](https://github.com/scverse/anndata) & [MuData](https://github.com/scverse/mudata) objects as input.
 
 <img src="https://raw.githubusercontent.com/saezlab/liana-py/main/docs/source/_static/abstract.png" width="700" align="center">
 
@@ -20,7 +20,7 @@ We welcome suggestions, ideas, and contributions! Please do not hesitate to cont
 A set of extensive vignettes can be found in the [LIANA+ documentation](https://liana-py.readthedocs.io/en/latest/).
 
 ## Decision Tree
-### Q: Does the data contain spatial coordinates?
+### Does the data contain spatial coordinates?
 #### Yes
 - **Q: Bivariate or unsupervised, multi-variate, and multi-view analysis?**
   - **Bivariate:**
@@ -41,6 +41,9 @@ A set of extensive vignettes can be found in the [LIANA+ documentation](https://
 - **Spatial:** [Integrating Multi-Modal Spatially-Resolved Technologies](https://liana-py.readthedocs.io/en/latest/notebooks/sma.html)
 - **Non-Spatial:** [Integrating Multi-Modal Single-Cell Technologies](https://liana-py.readthedocs.io/en/latest/notebooks/sc_multi.html)
 
+#### Infer Metabolite-mediated CCC from transcriptomics?
+[Non-spatial Data](https://liana-py.readthedocs.io/en/latest/notebooks/sc_multi.html#Metabolite-mediated-CCC-from-Transcriptomics-Data)
+
 ## API
 For further information please check LIANA's [API documentation](https://liana-py.readthedocs.io/en/latest/api.html).
 
@@ -50,4 +53,4 @@ Dimitrov D., Schäfer P.S.L, Farr E., Rodriguez Mier P., Lobentanzer S., Badia-i
 
 Dimitrov, D., Türei, D., Garrido-Rodriguez M., Burmedi P.L., Nagai, J.S., Boys, C., Flores, R.O.R., Kim, H., Szalai, B., Costa, I.G., Valdeolivas, A., Dugourd, A. and Saez-Rodriguez, J. Comparison of methods and resources for cell-cell communication inference from single-cell RNA-Seq data. Nat Commun 13, 3224 (2022). https://doi.org/10.1038/s41467-022-30755-0
 
-Similarly, please consider citing any of the methods and/or resources implemented in liana, that were particularly relevant for your research!
+Please also consider citing any of the methods and/or resources that were particularly relevant for your research!

docs/source/notebooks/basic_usage.ipynb

@@ -1634,14 +1634,14 @@
    "source": [
     "| Method                         | Magnitude                                                                                      | Specificity                                                                                  |\n",
     "|--------------------------------|------------------------------------------------------------------------------------------------|----------------------------------------------------------------------------------------------|\n",
-    "| CellPhoneDBv2                  | $$LRmean_{k,ij} = \\frac{L_{C_{i}} + R_{C_{j}}}{2}$$                                             | See ρ                                                                                        |\n",
+    "| [CellPhoneDBv2](https://www.nature.com/articles/s41596-020-0292-x)                  | $$LRmean_{k,ij} = \\frac{L_{C_{i}} + R_{C_{j}}}{2}$$                                             | See ρ                                                                                        |\n",
     "| Geometric Mean                 | $$LRgeometric.mean_{k,ij} = \\sqrt{L_{C_{i}} \\cdot R_{C_{j}}}$$                                 | See ρ                                                                                        |\n",
-    "| CellChat LR probabilities †      | $$LRprob_{k,ij} = \\frac{L^*_{C_{i}} \\cdot R^*_{C_{j}}}{Kh + L^*_{C_{i}} \\cdot R^*_{C_{j}}}$$ where Kh is a normalizing parameter (set to 0.5 by default) and L* & R* are aggregated using Tuckey's Trimean function (See below). | See ρ                                                                                        |\n",
-    "| SingleCellSignalR              | $$LRscore_{k,ij} = \\frac{\\sqrt{L_{C_{i}} R_{C_{j}}}}{\\sqrt{L_{C_{i}} R_{C_{j}}} + \\mu}$$ where $\\mu$ is the mean of the expression matrix M     | -                                                                                            |\n",
-    "| NATMI                          | $$LRproduct_{k,ij} = L_{C_{i}} R_{C_{j}}$$                                                    | $$SpecificityWeight_{k,ij} = \\frac{L_{C_{i}}}{\\sum^{n} L_{C_{i}}} \\cdot \\frac{R_{C_{j}}}{\\sum^{n} R_{C_{j}}}$$ |\n",
-    "| Connectome                     | $$LRproduct_{k,ij} = L_{C_{i}} R_{C_{j}}$$                                                    | $$LRz.mean_{k,ij} = \\frac{z_{L_{C_{i}}} + z_{R_{C_{j}}}}{2}$$   where z is the z-score of the expression matrix M|                                             |\n",
+    "| [CellChat's](https://www.nature.com/articles/s41467-021-21246-9) LR probabilities †      | $$LRprob_{k,ij} = \\frac{L^*_{C_{i}} \\cdot R^*_{C_{j}}}{Kh + L^*_{C_{i}} \\cdot R^*_{C_{j}}}$$ where Kh is a normalizing parameter (set to 0.5 by default) and L* & R* are aggregated using Tuckey's Trimean function (See below). | See ρ                                                                                        |\n",
+    "| [SingleCellSignalR](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7261168/)              | $$LRscore_{k,ij} = \\frac{\\sqrt{L_{C_{i}} R_{C_{j}}}}{\\sqrt{L_{C_{i}} R_{C_{j}}} + \\mu}$$ where $\\mu$ is the mean of the expression matrix M     | -                                                                                            |\n",
+    "| [NATMI](https://www.nature.com/articles/s41467-020-18873-z)                          | $$LRproduct_{k,ij} = L_{C_{i}} R_{C_{j}}$$                                                    | $$SpecificityWeight_{k,ij} = \\frac{L_{C_{i}}}{\\sum^{n} L_{C_{i}}} \\cdot \\frac{R_{C_{j}}}{\\sum^{n} R_{C_{j}}}$$ |\n",
+    "| [Connectome](https://www.nature.com/articles/s41598-022-07959-x)                     | $$LRproduct_{k,ij} = L_{C_{i}} R_{C_{j}}$$                                                    | $$LRz.mean_{k,ij} = \\frac{z_{L_{C_{i}}} + z_{R_{C_{j}}}}{2}$$   where z is the z-score of the expression matrix M|                                             |\n",
     "| LogFC‡                         | -                                                                                              | $$LRlog2FC_{k,ij} = \\frac{\\text{Log2FC}_{C_i,L} + \\text{Log2FC}_{C_j,R}}{2}$$                |\n",
-    "| ScSeqComm (intercellular scores only)   | $$LRinterscore_{k,ij} = \\text{min}(P(L_{Ci}), P(R_{Cj}))$$ $$P(X) = \\Phi\\left(\\frac{X - \\mu}{\\sigma / \\sqrt{n}}\\right)$$                              Where $\\Phi$ is the CDF of a normal distribution, μ is the mean, σ is the standard deviation, and n is the number of observations | -                                                                                            |\n",
+    "| [ScSeqComm](https://academic.oup.com/bioinformatics/article/38/7/1920/6511439) (intercellular scores only)   | $$LRinterscore_{k,ij} = \\text{min}(P(L_{Ci}), P(R_{Cj}))$$ $$P(X) = \\Phi\\left(\\frac{X - \\mu}{\\sigma / \\sqrt{n}}\\right)$$                              Where $\\Phi$ is the CDF of a normal distribution, μ is the mean, σ is the standard deviation, and n is the number of observations | -                                                                                            |\n",
     "| LIANA’s Consensus#             | Used flexibly to combine the Magnitude scores of the methods above. By default, uses all except the Geometric mean and CellChat, independently for magnitude and specificity scores. | Same as Magnitude Rank Aggregate but aggregates the specificity scores of different methods.                                                                                          |\n",
     "\n",
     "**Shared Notation:**\n",
@@ -1670,7 +1670,7 @@
     "\n",
     "where m is the number of ranked score vectors, n is the length of each score vector (number of interactions), rankij is the rank of the j-th element (interaction) in the i-th score rank vector, and max(ranki) is the maximum rank in the i-th rank vector.\n",
     "\n",
-    "For each normalised rank vector r, we then ask how probable it is to obtain rnull(k)<= r(k), where rnull(k) is a rank vector generated under the null hypothesis. The RobustRankAggregate method 9 expresses the probability rnull(k)<= r(k) as k,n(r) through a beta distribution. This entails that we obtain probabilities for each score vector r as:\n",
+    "For each normalised rank vector r, we then ask how probable it is to obtain rnull(k)<= r(k), where rnull(k) is a rank vector generated under the null hypothesis. The [RobustRankAggregate](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3278763/) method expresses the probability rnull(k)<= r(k) as k,n(r) through a beta distribution. This entails that we obtain probabilities for each score vector r as:\n",
     "\n",
     "$$p(r) = \\underset{1, ..., n}{min} \\beta_k,_n(r) * n$$\n",
     "\n",

docs/source/notebooks/prior_knowledge.ipynb

@@ -11,7 +11,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "LIANA+ typically relies heavily on prior knowledge to infer intercellular communication and the intracellular signaling pathways that are activated in response to communication. This notebook provides a brief overview of the prior knowledge typically used by LIANA+. "
+    "LIANA+ (typically) relies heavily on prior knowledge to infer intercellular communication and the intracellular signaling pathways that are activated in response to communication. This notebook provides a brief overview of the prior knowledge typically used by LIANA+. "
    ]
   },
   {

docs/source/notebooks/sc_multi.ipynb

@@ -516,7 +516,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "Recently, tools such as NeuronChat, scConnect, Cellinker, and CellPhoneDBv5 have proposed simple methods, such as enrichment, expression average, among others, to infer metabolite-mediated CCC events from transcriptomics data. Similarly, we can use LIANA+ to infer metabolite-mediated CCC events from transcriptomics data, as [described in the MetalinksDB manuscript](https://www.biorxiv.org/content/10.1101/2023.12.30.573715v1.abstract).\n",
+    "Recently, tools such as [NeuronChat](https://www.nature.com/articles/s41467-023-36800-w), [MEBOCOST](https://www.biorxiv.org/content/10.1101/2022.05.30.494067v2.abstract), [scConnect](https://academic.oup.com/bioinformatics/article/37/20/3501/6273571), [Cellinker](https://academic.oup.com/bioinformatics/article/37/14/2025/6104823), and [CellPhoneDBv5](https://arxiv.org/abs/2311.04567) have proposed approaches, such as enrichment, expression average, among others, to infer metabolite-mediated CCC events from transcriptomics data. Similarly, we can use LIANA+ to infer metabolite-mediated CCC events from transcriptomics data, as [described in the MetalinksDB manuscript](https://academic.oup.com/bib/article/25/4/bbae347/7717953).\n",
     "\n",
     "Briefly, we use a univariate linear regression model to estimate metabolite abundances for each cell. To do so, we make use of production-degradation enzyme prior knowledge to infer the metabolite abundances. Optionally, we also take transporters into account. We then use these inferred metabolite abundances to infer metabolite-mediated CCC events.\n",
     "\n",

docs/source/release_notes.rst

@@ -2,6 +2,7 @@ Changelog
 =============
 
 1.4.0 (02.09.2024)
+-------------------------------------------------
 
 - Now published at Nat Cell Bio.
 
@@ -13,6 +14,7 @@ This might result in minor differences of how many interactions are considered p
 - Changed ``max_neighbours`` in ``li.ut.spatial_neighbors`` to be a fixed number (default=100), rather than a fraction of the spots as this was making RAM explode for large spatial formats.
 
 1.3.0 (12.07.2024)
+-------------------------------------------------
 
 - Minor improvements to documentation, specifically changed to the furo theme. Resolved issues with latex not being rendered and plot sizes being off.
 
@@ -183,6 +185,7 @@ Essentially, it diminishes the effect of spatial proximity on the score, while s
 - Bumped CORNETO version and it's now installed via PyPI.
 
 1.0.0a2 (19.09.2023)
+-------------------------------------------------
 
 - Interactions names in ``tileplot`` and ``dotplot`` will now be sorted according to ``orderby`` when used; related to #55