There are several basic questions of developmental biology :

1. What classes of cells are present at each stage?

2. For the cells in each class, what was their origin at earlier stages, what are their potential fates at later stages, and what is the actual outcome of a given cell?

3. To what extent are events along a path synchronous or asynchronous?

4. What are the genetic regulatory programs that control each path?

5. What are the intercellular interactions between classes of cells?

6. How deterministic or stochastic is the process—that is: if, and how early, does it become determined that a particular cell or an entire cell class is destined to a specific fate?

The resolution level of single cell can be used to analyze the early development process of embryo , cell atlase of the development landscape represents an important step towards understanding organogenesis, which can contains continuous time and spatial information of the cell, here we focus on trajectory inference of organogenesis: #1choose distinguish factors in cell phase transition
1.1based on prior knowledge such as capture times (DeLorean) or switch-like marker genes (Ouija);
1.2select gene with high depersion across cells;
1.3ordering based on genes that differ between clusters and define a cell's progress

Sum: It is difficult to reconstruct the development trajectory of single cells due to the lack of known reference genes for the transition point of biological state, or when reference genes are insufficient to span the entire observation period window.

2 Modeling specific types of biological processes such as branching processes in differentiation (multiple methods) or cyclic processes (Oscope) to infer pseudotime

2.1Laleh develop a pseudotime measure call diffusion pseudotime (DPT), which measures transitions between cells using diffusion-like random walks. DPT implementations make it possible to reconstruct the developmental progression of cells and identify transient or metastable states, branching decisions and differentiation endpoints. (Laleh et al.,2016).

2.2Saelens et al. 2018 performed a comprehensive evaluation of 29 different single-cell trajectory inference methods and discuss the different types of algorithms in more detail. They benchmark both quantitative performance and assess software quality.

Sum: During development, the state of the cells is constantly changing and not synchronous. ScRNA-Seq can separate the cells in the middle state of the process, and use algorithms to learn the expression patterns of all cells to arrange them into their respective developmental trajectories.

we focus on trajectory inference of organogenesis:
This protocol is to sample at even interval time points to obtain a sufficient number of cells of single-cell transcriptome data, then it can be assumed in the analysis that these single cells are in a continuously changing state, and have cells in both continuous time and spatial information. ( Blanca et al.,2019)Thus, all of these cells can be arranged by pseudo-time and other path reduction algorithms (Trapnell et al., 2014). To validate this reconstructed development trajectory, they focused on reconstructing skeletal myogenesis, which comprises distinct mesodermal lineages, the trajectory is consistent with the hypothesis that different mesodermal lineages use distinct factors to converge on a core program of muscle genes. (Cao et al.,2019) Monocle has high accuracy and can be easily transplanted into other types of cell data analysis. I tried the sample code of the monocle test data. The demo of the official website is very detailed, but it is more difficult for me to understand the algorithm. The transcripts of tens of thousands of single cells are sequenced, and it is therefore hoped that a single cell of various states will be continuously distributed in the experiment. It can increase the temporal resolution of transcriptome dynamics using single-cell RNA-Seq data collected at multiple time points.

along with pseudotime + spational transcription

3.spatial transcriptomic analysis of tissue -Landmark gene with specific loci expression is known.

View ORCID ProfileWouter Saelens, View ORCID ProfileRobrecht Cannoodt, View ORCID ProfileHelena Todorov, View ORCID ProfileYvan Saeys,A comparison of single-cell trajectory inference methods: towards more accurate and robust tools

以作者把Optimal Transport (OT)的算法，应用到了时间序列的单细胞转录组数据来探索发育的过程。当然，表现很好的啦，揭示了重编程的分子机理。 几大发现如下：

·(1) identifying alternative cell fates, including senescence, apoptosis, neural identity, and placental identity;

· (2) quantifying the portion of cells in each state at each time point;

· (3) inferring the probable origin(s) and fate(s) of each cell and cell class at each time point;

· (4) identifying early molecular markers associated with eventual fates;

· (5) using trajectory information to identify transcription factors (TFs) associated with the activation of different expres

部分参考：
作者：小梦游仙境 链接：https://www.jianshu.com/p/2911b5107605 来源：简书 简书著作权归作者所有，任何形式的转载都请联系作者获得授权并注明出处。