The new development of RNALocate.

1. Expand data sources and coverage of species, subcellular localizations and RNA types.

2. Add comprehensive expression analysis of mRNAs/lncRNAs/miRNAs from high-throughput sequencing data of subcellular localization, in particular dynamic changes of RNA localization under different conditions.

3. Provide a new multi-label RNA localization prediction tool covering the subcellular localization of 7 types of RNA, including mRNA, lncRNA, miRNA, piRNA, snoRNA, snRNA and circRNA, across 11 different subcellular compartments.

4. Redesigned database based on the Django Model-Template-View (MTV) framework.

5. Add filter tool in result page of search and browse.

Fig 1-1:

1. Main functions of the database are provided in menu bar form (boxed in red).

2. Quick Search be used to quickly retrieve experimental validation data and RNA-seq analysis data.

3. Users can quickly access to use features of RNALocate.

4. Contents of RNALocate updates.

5. Other databases contributed by our group.

6. Cite information.

The Exact search page is displayed in Fig 2-1:

1. Carefully select a keyword type: Four choices are provided.

2. Enter a keyword corresponding to selected keyword type.

3. Five categories provided to filter results: Category, Species, Data Source, Minimum Score and Maximum Score.

4. Use NCBI Gene / Ensembl / EnsemblRapid / LncBook / miRBase / circBase / circBank / piRBase / RNAcentral / exoRBase / Gene Ontology to normalize your input information.

The Fuzzy search page is displayed in Fig 2-2:

1. Select the category of your keyword.

2. Enter a keyword corresponding to selected.

3. Choose the match entries what you want.

The Batch search page is displayed in Fig 2-3:

1. Select a type for your keywords.

2. Enter the keywords or upload a file.

This tutorial is as follows.

1. First we have to choose the type of keyword. There are four keyword types in our search as the picture shows. In this example, we choose ‘RNA Symbol’ as the keyword type.

2. Next, we enter the keyword according to the keyword type selected in the previous step. In this example, we choose 'ACTB' as the keyword.

3. Then select the category for the keyword you entered. In this example, we choose 'mRNA' as the category of the keyword 'ACTB'.

4. Then select the species for the keyword you entered. In this example, we choose 'Homo sapiens' as the species of the keyword 'ACTB'.

5. You can choose the type of data source as the filter. In this example, we want query the associations detected by experimental evidence, so we choose 'All Experimental Evidence'.

6. We provide a score for each association. The greater the value, the higher the credibility. To filter low-confidence associations, in this example, we choose the 0.5 as the minimum score and 1.0 as the maximum score.

7. With all the filters above, we can click 'Search' to query the result.

8. After several seconds, the result will appear. All the associations are represented in the table format.

In the result page, all entries are listed with basic information including RNA symbols,RNA categories, species, tissue/cell lines, localizations and scores.

1. Use the filter option on the left to further screen search results by score, species, RNA category and localization.

2. Choose keyword on the top to sort all search results.

3. Click to download the results page.

4. Click to turn the page.

5. Click to link to detail page.

In the detail page, you can get information including RNALocate ID, confidence score, RNA information, localization information, homology information, other subcellular localization, RNA disease information and RNA interaction information.

Fig 5-1:

We redesigned the RNALocate ID. Take RL-E-m-0034508 as an example, where 'RL' stands for RNALocate database, 'E' stands for the data type of experimental verification, while 'P' stands for prediction data. 'm' stands for the RNA type is mRNA, and similarly, 'mi' stands for miRNA, 'lnc' stands for RNA type is lncRNA, 'circ' stands for circRNA, 'pi' stands for piRNA, 'sno' stands for snoRNA, 't' stands for tRNA and 34508 is the unique ID number.

Fig 5-2:

1. For RNA information, RNA name from literature, RNA ID, RNA category, Species and Alias of each entry. At the same time, we give the location information of miRNA/circRNA/lncRNA (from LncBook).

2. Click RNA ID to see its basic description in NCBI Gene / Ensembl / EnsemblRapid / LncBook / miRBase / circBase / circBank / piRBase / RNAcentral / exoRBase database.

3. Click Show button to see its sequence of the RNA (detail in Fig 5-3).

Fig 5-4:

1. Related subcellular localization information .

2. Click Localization to see its basic description in Gene Ontology database.

Fig 5-5:

1. The data source, localization in paper, tissue or cell line, method, description and pubmed date of each entry.

2. Click pubmed ID to see description in detail.

Fig 5-6:

1. The association of homology information has been provided.

2. Click Go to button to search in RNALocate.

Fig 5-7:

1. The other subcellular localization of this RNA has been provided.

2. Click RNALocate ID to see detail information of this entry.

Fig 5-8:

Disease information for each RNA in our database is provided. Click the Link to see more details about them.

Fig 5-9:

Interactions for each RNA in our database are provided. Click the Link to search for more details about them and Click the RNA2 to obtain the corresponding RNA-localization information.

Fig 5-10:

1. Name of prediction algorithm.

2. Score of prediction algorithm.

3. Principle of prediction algorithm.

In the browse page, you can click each node to see results.

1. 'Localization' displays all entries as long as the current selected localization is involved.

2. 'RNA Category' indicates all types of RNA in RNALocate.

3. 'Species' displays all entries as long as the organism matches the condition.

4. 'Prediction Algorithm' displays all entries as long as the Algorithm matches the condition.

‘RNA-seq Analysis’ performs comprehensive analysis and visualization of RNA-seq data of more than 850 subcellular localization-associated samples from Gene Expression Omnibus (GEO), Sequence Read Archive (SRA), ArrayExpress and European Nucleotide Archive (ENA).

1. RNALocate normalized the expression level by TPM. If the number of samples is not less than two and the TPM of at least two samples is greater than 0, then the mean and median are calculated as the RNA expression in the subcellular localization under the condition.

2. RNALocate uses DESeq2/edgeR to identify differentially expressed RNAs in subcellular localization or subcellular-specific localization under different conditions with FDR (False Discovery Rate) value < 0.05 and FC (Fold Change) > 1.2 as thresholds.

3. RNALocate performs functional enrichment analysis on differentially expressed mRNA/miRNAs (target genes predicted by miRWalk) and sets FDR (False Discovery Rate) value <= 0.05 and FC (Fold Change) > 1.2 as the threshold.

4. RNALocate employed RNAfold to predict secondary structure and calculate the minimum free energy (MFE), which was then normalized (NMFE) based on sequence length.

5. RNALocate used RBPmap to predict potential RBPs binding to RNA, and the parameters were set to be Stringency level (high stringency)

6. RNALocate calculate the phastCons conservation score for each RNA.

Search for dataset (From dataset)

Fig 7-1

1. Filter datasets by source, species, tissue/cell line and location, and choose a dataset according to which subcellular localization users are interested in.

2. Click "More" to see detailed information about this dataset.

Fig 7-2:

Dataset number, status, species, source, subcellular localization, pubmed ID, title, citation and experimental design were provided.

Fig 7-3:

Each sample ID, bioproject, localization, condition (including treatment, time, genotype and library type, etc.) and label were provided.

Fig 7-4:

The numbers of RNA-enriched locations under various conditions have been provided. Users can know the number of RNA-enriched locations by selecting RNA type and condition. The bar chart on the left shows the number of RNAs corresponding to different subcellular localizations, while the donut chart on the right illustrates the detailed distribution shown in the left bar chart, indicating which combinations of subcellular localizations and their corresponding amounts of RNA are represented under each set.

1. Select a type of RNA.

2. Select the condition. Condition All represents the union of all conditions.

3. Hovering the cursor over the bar chart or donut chart will display the details.

4. Click on any part of the donut chart to get the corresponding gene information.

Fig 7-5:

Pairwise differential analysis of different subcellular localizations is performed by DESeq2/edgeR under the same conditions, with FDR (False Discovery Rate) value < 0.05 and FC (Fold Change) > 1.2 as the threshold. Meanwhile, RNALocate performs functional enrichment analysis on differentially expressed mRNA/miRNAs (target genes predicted by miRWalk) and sets FDR (False Discovery Rate) value <= 0.05 and FC (Fold Change) > 1.2 as the threshold.

1. Select a type of RNA.

2. Select location for which differential analysis will be performed.

3. Select the condition.

4. Select differential method.

5. Select enrichment method.

6. The volcano plot corresponding to the above 5 conditions.

7. The top 200 upregulated RNAs in the volcano plot.

8. The most significantly related pathways or functions of upregulated RNAs in the volcano plot, usually the top 20.

Fig 7-6:

Under the same conditions, the subcellular-specific RNAs are the intersection of differentially upregulated RNAs between a given subcellular localization and all other subcellular localizations.

1. Select a type of RNA.

2. Select differential method.

3. Hovering the cursor over the barplot will display the details.

4. Click on any part of the barplot to get the corresponding RNAs information and expression information.

5. Click on any part of the barplot to get the histogram about NMFE (normalized minimum free energy) for these RNAs.

6. Click on any part of the barplot to get the barplot about predicted top 20 RBP for these RNAs.

7. Click on any part of the barplot to get the histogram about phastCons conservation scores for these RNAs.

Fig 7-7:

Dynamic change of subcellular-specific RNAs under different conditions. NS: no specific; RNA localization was not specific, i.e., RNA localized to at least two subcellular fractions after differentiation.

1. Select a type of RNA.

2. Select differential method.

3. Hovering the cursor over the sankey diagram will display the details.

4. Click on any part of the sankey diagram to get the corresponding gene information and localization information under various conditions in this dataset.

Fig 7-8:

RNALocate performs functional enrichment analysis on subcellular-specific mRNA/miRNAs (target genes predicted by miRWalk)and shows some of the most significantly related pathways or functions, setting FDR (False Discovery Rate) value <= 0.05, usually the top 20.

1. Select a type of RNA.

2. Select localization.

3. Select the condition. Condition All represents the union of all conditions.

4. Select differential method.

5. Select enrichment method.

Fig 7-9:

Localization ratios of subcellular-specific RNAs under different conditions, with FDR (False Discovery Rate) value <= 0.05 and FC (Fold Change) > 1.2 as the threshold.

1. Select a type of RNA.

2. Select a RNA symbol.

Search for RNA symbol/RNA ID (From RNA page)

Fig 7-10:

1. Use RNA symbol or RNA ID as a keyword to search.

2. Select species.

Firstly, you can get basic information about each RNA in the detail page.

Second, you can switch to the localization from literature mining search result page.

Third, you can see the subcellular localization of RNA in different RNA-seq dataset.

Fig 7-11:

1. Click RNA ID to see its description in official database.

2. Click Detail button to search the localization of this RNA from literature mining.

Fig 7-12:

Barplot: Subcellular localization and expression of RNA under different conditions.

Table: Relative ratio of RNA expression in different locations and conditions. The greater the relative ratio, the higher the abundance of RNA in that particular location and under those specific conditions.

Note: The calculation method of relative ratio and the dataset number.

1. Select dataset.

2. Select TPM mean or median.

3. Hovering the cursor over the barplot will display the details.

4. Click the dataset number to jump to the corresponding dataset.

The Tool provides a new multi-label RNA subcellular localization prediction tool in our database, including mRNA, lncRNA, miRNA, piRNA, snoRNA, snRNA and circRNA.

Fig 8-1: Prediction tool page.

1. Choose RNA type.

2. Directly input one or more mRNA/lncRNA/miRNA/piRNA/snoRNA/snRNA/circRNA sequences or upload a .fasta file with fasta format.

3. Click Run button.

Fig 8-2: Prediction tool result page.

Display predicted results and support downloads. The subcellular localizations with the top three scores will be displayed. If the predicted result for an RNA is chromatin--(nucleus), then the RNA is located in chromatin and also in the nucleus.

In RNALocate, the RNA-subcellular localization associations are collected from different types of resources under one common framework, including experimental and prediction evidence. In principle, we assume that:

1. Experimental evidence should contribute more significantly to the confidence score than predictive evidence;

2. Strong experimental evidence should provide more reliable evidence than weak experimental evidence;

3. RNA-subcellular localization associations supported by more evidence should be given significantly higher confidence scores than those supported by less evidence.

Similar to RNADisease, according to the evidence types and number of evidence resources, we calculate the confidence score (S) for each RNA-subcellular localization association as follows:

where i is the evidence type (s: strong experimental evidence, w: weak experimental evidence, p: computational prediction method), x is the number of evidence resources, we set weight factor Ws, Ww and Wp to 0.95, 0.45, and 0.1, respectively (if x=0, we set weight factor Wi to 0).

Integration of source databases which use different RNA and subcellular localization naming conventions is challenging. To ensure maximal connectivity of data, we transform each RNA and subcellular localization name found in the input sources to the appropriate naming convention.

1. For miRNA, we use miRBase ID, miRBase Accession, NCBI Gene Symbol and Entrez ID.

2. For circRNA, we use circBase ID, circBank ID and exoRBase ID.

3. For piRNA, we use piRBase Name, piRBase Accession and RNAcentral identifier.

4. For lncRNA, we use LncBook ID, NCBI Gene Symbol, Entrez ID, Ensembl Gene Stable ID and RNAcentral identifier.

5. For others, we use NCBI Gene Symbol, Entrez ID, Ensembl Gene Stable ID, EnsemblRapid Gene Stable ID and RNAcentral identifier.

6. For subcellular localization, we use cellular component of Gene Ontology.

7. For species, we normalized organism names according to NCBI Taxonomy Database.